C | Domain-Specific Machine Learning

Controllable text-to-image (T2I) diffusion models have shown impressive performance in generating high-quality visual content through the incorporation of various conditions. Current methods, however, exhibit limited performance when guided by skeleton human poses, especially in complex pose conditions such as side or rear perspectives of human figures. To address this issue, we present Stable-Pose, a novel adapter model that introduces a coarse-to-fine attention masking strategy into a vision Transformer (ViT) to gain accurate pose guidance for T2I models. Stable-Pose is designed to adeptly handle pose conditions within pre-trained Stable Diffusion, providing a refined and efficient way of aligning pose representation during image synthesis. We leverage the query-key self-attention mechanism of ViTs to explore the interconnections among different anatomical parts in human pose skeletons. Masked pose images are used to smoothly refine the attention maps based on target pose-related features in a hierarchical manner, transitioning from coarse to fine levels. Additionally, our loss function is formulated to allocate increased emphasis to the pose region, thereby augmenting the model’s precision in capturing intricate pose details. We assessed the performance of Stable-Pose across five public datasets under a wide range of indoor and outdoor human pose scenarios. Stable-Pose achieved an AP score of 57.1 in the LAION-Human dataset, marking around 13% improvement over the established technique ControlNet.

Topological correctness, i.e., the preservation of structural integrity and specific characteristics of shape, is a fundamental requirement for medical imaging tasks, such as neuron or vessel segmentation. Despite the recent surge in topology-aware methods addressing this challenge, their real-world applicability is hindered by flawed benchmarking practices. In this paper, we identify critical pitfalls in model evaluation that include inadequate connectivity choices, overlooked topological artifacts in ground truth annotations, and inappropriate use of evaluation metrics. Through detailed empirical analysis, we uncover these issues’ profound impact on the evaluation and ranking of segmentation methods. Drawing from our findings, we propose a set of actionable recommendations to establish fair and robust evaluation standards for topology-aware medical image segmentation methods.

Digital pathology offers a groundbreaking opportunity to transform clinical practice in histopathological image analysis, yet faces a significant hurdle: the substantial file sizes of pathological Whole Slide Images (WSI). While current digital pathology solutions rely on lossy JPEG compression to address this issue, lossy compression can introduce color and texture disparities, potentially impacting clinical decision-making. While prior research addresses perceptual image quality and downstream performance independently of each other, we jointly evaluate compression schemes for perceptual and downstream task quality on four different datasets. In addition, we collect an initially uncompressed dataset for an unbiased perceptual evaluation of compression schemes. Our results show that deep learning models fine-tuned for perceptual quality outperform conventional compression schemes like JPEG-XL or WebP for further compression of WSI. However, they exhibit a significant bias towards the compression artifacts present in the training data and struggle to generalize across various compression schemes. We introduce a novel evaluation metric based on feature similarity between original files and compressed files that aligns very well with the actual downstream performance on the compressed WSI. Our metric allows for a general and standardized evaluation of lossy compression schemes and mitigates the requirement to independently assess different downstream tasks. Our study provides novel insights for the assessment of lossy compression schemes for WSI and encourages a unified evaluation of lossy compression schemes to accelerate the clinical uptake of digital pathology.

3D Gaussian Splatting has recently gained traction for its efficient training and real-time rendering. While the vanilla Gaussian Splatting representation is mainly designed for view synthesis, more recent works investigated how to extend it with scene understanding and language features. However, existing methods lack a detailed comprehension of scenes, limiting their ability to segment and interpret complex structures. To this end, We introduce SuperGSeg, a novel approach that fosters cohesive, context-aware scene representation by disentangling segmentation and language field distillation. SuperGSeg first employs neural Gaussians to learn instance and hierarchical segmentation features from multi-view images with the aid of off-the-shelf 2D masks. These features are then leveraged to create a sparse set of what we call Super-Gaussians. Super-Gaussians facilitate the distillation of 2D language features into 3D space. Through Super-Gaussians, our method enables high-dimensional language feature rendering without extreme increases in GPU memory. Extensive experiments demonstrate that SuperGSeg outperforms prior works on both open-vocabulary object localization and semantic segmentation tasks.

Image registration is fundamental in medical imaging applications, such as disease progression analysis or radiation therapy planning. The primary objective of image registration is to precisely capture the deformation between two or more images, typically achieved by minimizing an optimization problem. Due to its inherent ill-posedness, regularization is a key component in driving the solution toward anatomically meaningful deformations. A wide range of regularization methods has been proposed for both conventional and deep learning-based registration. However, the appropriate application of regularization techniques often depends on the specific registration problem, and no one-fits-all method exists. Despite its importance, regularization is often overlooked or addressed with default approaches, assuming existing methods are sufficient. A comprehensive and structured review remains missing. This review addresses this gap by introducing a novel taxonomy that systematically categorizes the diverse range of proposed regularization methods. It highlights the emerging field of learned regularization, which leverages data-driven techniques to automatically derive deformation properties from the data. Moreover, this review examines the transfer of regularization methods from conventional to learning-based registration, identifies open challenges, and outlines future research directions. By emphasizing the critical role of regularization in image registration, we hope to inspire the research community to reconsider regularization strategies in modern registration algorithms and to explore this rapidly evolving field further.

Biophysical modeling of brain tumors has emerged as a promising strategy for personalizing radiotherapy planning by estimating the otherwise hidden distribution of tumor cells within the brain. However, many existing state-of-the-art methods are computationally intensive, limiting their widespread translation into clinical practice. In this work, we propose an efficient and direct method that utilizes soft physical constraints to estimate the tumor cell concentration from preoperative MRI of brain tumor patients. Our approach optimizes a 3D tumor concentration field by simultaneously minimizing the difference between the observed MRI and a physically informed loss function. Compared to existing state-of-the-art techniques, our method significantly improves predicting tumor recurrence on two public datasets with a total of 192 patients while maintaining a clinically viable runtime of under one minute - a substantial reduction from the 30 minutes required by the current best approach. Furthermore, we showcase the generalizability of our framework by incorporating additional imaging information and physical constraints, highlighting its potential to translate to various medical diffusion phenomena with imperfect data.

Registering pre-operative modalities, such as magnetic resonance imaging or computed tomography, to ultrasound images is crucial for guiding clinicians during surgeries and biopsies. Recently, deep-learning approaches have been proposed to increase the speed and accuracy of this registration problem. However, all of these approaches need expensive supervision from the ultrasound domain. In this work, we propose a multitask generative framework that needs weak supervision only from the pre-operative imaging domain during training. To perform a deformable registration, the proposed framework translates a magnetic resonance image to the ultrasound domain while preserving the structural content. To demonstrate the efficacy of the proposed method, we tackle the registration problem of pre-operative 3D MR to transrectal ultrasonography images as necessary for targeted prostate biopsies. We use an in-house dataset of 600 patients, divided into 540 for training, 30 for validation, and the remaining for testing. An expert manually segmented the prostate in both modalities for validation and test sets to assess the performance of our framework. The proposed framework achieves a 3.58 mm target registration error on the expert-selected landmarks, 89.2% in the Dice score, and 1.81 mm 95th percentile Hausdorff distance on the prostate masks in the test set. Our experiments demonstrate that the proposed generative model successfully translates magnetic resonance images into the ultrasound domain. The translated image contains the structural content and fine details due to an ultrasound-specific two-path design of the generative model. The proposed framework enables training learning-based registration methods while only weak supervision from the pre-operative domain is available.

This study investigates the predictive capability of radiomics in determining programmed cell death ligand 1 (PD-L1) expression (>=1%) status in non-small cell lung cancer (NSCLC) patients using a newly collected [18F]FDG PET/CT dataset. We aimed to replicate and validate the radiomics-based machine learning (ML) model proposed by Zhao et al. [2] predicting PD-L1 status from PET/CT-imaging.
An independent cohort of 254 NSCLC patients underwent [18F]FDG PET/CT imaging, with primary tumor segmentation conducted using lung tissue window (LTW) and more conservative soft tissue window (STW) methods. Radiomics models (“Rad-score” and “complex model”) and a clinical-stage model from Zhao et al. were evaluated via 10-fold cross-validation and AUC analysis, alongside a benchmark-study comparing different ML-model pipelines. Clinicopathological data were collected from medical records.
On our data, the Rad-score model yielded mean AUCs of 0.593 (STW) and 0.573 (LTW), below Zhao et al.’s 0.761. The complex model achieved mean AUCs of 0.505 (STW) and 0.519 (LTW), lower than Zhao et al.’s 0.769. The clinical model showed a mean AUC of 0.555, below Zhao et al.’s 0.64. All models performed significantly lower than Zhao et al.’s findings. Our benchmark study on four ML pipelines revealed consistently low performance across all configurations.
Our study failed to replicate original findings, suggesting poor model performance and questioning predictive value of radiomics features in classifying PD-L1 expression from PET/CT imaging. These results highlight challenges in replicating radiomics-based ML models and stress the need for rigorous validation

Meningeal lymphatic vessels (MLVs) are responsible for the drainage of waste products from the human brain. An impairment in their functionality has been associated with aging as well as brain disorders like multiple sclerosis and Alzheimer’s disease. However, MLVs have only recently been described for the first time in magnetic resonance imaging (MRI), and their ramified structure renders manual segmentation particularly difficult. Further, as there is no consistent notion of their appearance, human-annotated MLV structures contain a high inter-rater variability that most automatic segmentation methods cannot take into account. In this work, we propose a new rater-aware training scheme for the popular nnU-Net model, and we explore rater-based ensembling strategies for accurate and consistent segmentation of MLVs. This enables us to boost nnU-Net’s performance while obtaining explicit predictions in different annotation styles and a rater-based uncertainty estimation. Our final model, MLV2-Net, achieves a Dice similarity coefficient of 0.806 with respect to the human reference standard. The model further matches the human inter-rater reliability and replicates age-related associations with MLV volume.

Topological correctness plays a critical role in many image segmentation tasks, yet most networks are trained using pixel-wise loss functions, such as Dice, neglecting topological accuracy. Existing topology-aware methods often lack robust topological guarantees, are limited to specific use cases, or impose high computational costs. In this work, we propose a novel, graph-based framework for topologically accurate image segmentation that is both computationally efficient and generally applicable. Our method constructs a component graph that fully encodes the topological information of both the prediction and ground truth, allowing us to efficiently identify topologically critical regions and aggregate a loss based on local neighborhood information. Furthermore, we introduce a strict topological metric capturing the homotopy equivalence between the union and intersection of prediction-label pairs. We formally prove the topological guarantees of our approach and empirically validate its effectiveness on binary and multi-class datasets. Our loss demonstrates state-of-the-art performance with up to fivefold faster loss computation compared to persistent homology methods.

Employing pre-trained Large Language Models (LLMs) has become the de facto standard in Natural Language Processing (NLP) despite their extensive data requirements. Motivated by the recent surge in research focused on training LLMs with limited data, particularly in low-resource domains and languages, this paper surveys recent transfer learning approaches to optimize model performance in downstream tasks where data is scarce. We first address initial and continued pre-training strategies to better leverage prior knowledge in unseen domains and languages. We then examine how to maximize the utility of limited data during fine-tuning and few-shot learning. The final section takes a task-specific perspective, reviewing models and methods suited for different levels of data scarcity. Our goal is to provide practitioners with practical guidelines for overcoming the challenges posed by constrained data while also highlighting promising directions for future research.

Ultrasound is widely used in medical diagnostics allowing for accessible and powerful imaging but suffers from resolution limitations due to diffraction and the finite aperture of the imaging system, which restricts diagnostic use. The impulse function of an ultrasound imaging system is called the point spread function (PSF), which is convolved with the spatial distribution of reflectors in the image formation process. Recovering high-resolution reflector distributions by removing image distortions induced by the convolution process improves image clarity and detail. Conventionally, deconvolution techniques attempt to rectify the imaging system’s dependent PSF, working directly on the radio-frequency (RF) data. However, RF data is often not readily accessible. Therefore, we introduce a physics-based deconvolution process using a modeled PSF, working directly on the more commonly available B-mode images. By leveraging Implicit Neural Representations (INRs), we learn a continuous mapping from spatial locations to their respective echogenicity values, effectively compensating for the discretized image space. Our contribution consists of a novel methodology for retrieving a continuous echogenicity map directly from a B-mode image through a differentiable physics-based rendering pipeline for ultrasound resolution enhancement. We qualitatively and quantitatively evaluate our approach on synthetic data, demonstrating improvements over traditional methods in metrics such as PSNR and SSIM. Furthermore, we show qualitative enhancements on an ultrasound phantom and an in-vivo acquisition of a carotid artery.

In radiation therapy (RT), an accurate delineation of the regions of interest (ROI) and organs at risk (OAR) allows for a more targeted irradiation with reduced side effects. The current clinical workflow for combined MR-linear accelerator devices (MR-linacs) requires the acquisition of a planning MR volume (MR-P), in which the ROI and OAR are accurately segmented by the clinical team. These segmentation maps (S-P) are transferred to the MR acquired on the day of the RT fraction (MR-Fx) using registration, followed by time-consuming manual corrections. The goal of this paper is to enable accurate automatic segmentation of MR-Fx using S-P without clinical workflow disruption. We propose a novel UNet-based architecture, CloverNet, that takes as inputs MR-Fx and S-P in two separate encoder branches, whose latent spaces are concatenated in the bottleneck to generate an improved segmentation of MP-Fx. CloverNet improves the absolute Dice Score by 3.73% (relative +4.34%, p<0.001) when compared with conventional 3D UNet. Moreover, we believe this approach is potentially applicable to other longitudinal use cases in which a prior segmentation of the ROI is available.

Generally, the small size of public medical imaging datasets coupled with stringent privacy concerns, hampers the advancement of data-hungry deep learning models in medical imaging. This study addresses these challenges for 3D cardiac MRI images in the short-axis view. We propose Latent Diffusion Models that generate synthetic images conditioned on medical attributes, while ensuring patient privacy through differentially private model training. To our knowledge, this is the first work to apply and quantify differential privacy in 3D medical image generation. We pre-train our models on public data and finetune them with differential privacy on the UK Biobank dataset. Our experiments reveal that pre-training significantly improves model performance, achieving a Fréchet Inception Distance (FID) of 26.77 at ϵ=10, compared to 92.52 for models without pre-training. Additionally, we explore the trade-off between privacy constraints and image quality, investigating how tighter privacy budgets affect output controllability and may lead to degraded performance. Our results demonstrate that proper consideration during training with differential privacy can substantially improve the quality of synthetic cardiac MRI images, but there are still notable challenges in achieving consistent medical realism.

Surgical data science (SDS) is a field that analyzes patient data before, during, and after surgery to improve surgical outcomes and skills. However, surgical data is scarce, heterogeneous, and complex, which limits the applicability of existing machine learning methods. In this work, we introduce the novel task of future video generation in laparoscopic surgery. This task can augment and enrich the existing surgical data and enable various applications, such as simulation, analysis, and robot-aided surgery. Ultimately, it involves not only understanding the current state of the operation but also accurately predicting the dynamic and often unpredictable nature of surgical procedures. Our proposed method, VISAGE (VIdeo Synthesis using Action Graphs for Surgery), leverages the power of action scene graphs to capture the sequential nature of laparoscopic procedures and utilizes diffusion models to synthesize temporally coherent video sequences. VISAGE predicts the future frames given only a single initial frame, and the action graph triplets. By incorporating domain-specific knowledge through the action graph, VISAGE ensures the generated videos adhere to the expected visual and motion patterns observed in real laparoscopic procedures. The results of our experiments demonstrate high-fidelity video generation for laparoscopy procedures, which enables various applications in SDS.

Graph-based holistic scene representations facilitate surgical workflow understanding and have recently demonstrated significant success. However, this task is often hindered by the limited availability of densely annotated surgical scene data. In this work, we introduce an end-to-end framework for the generation and optimization of surgical scene graphs on a downstream task. Our approach leverages the flexibility of graph-based spectral clustering and the generalization capability of foundation models to generate unsupervised scene graphs with learnable properties. We reinforce the initial spatial graph with sparse temporal connections using local matches between consecutive frames to predict temporally consistent clusters across a temporal neighborhood. By jointly optimizing the spatiotemporal relations and node features of the dynamic scene graph with the downstream task of phase segmentation, we address the costly and annotation-burdensome task of semantic scene comprehension and scene graph generation in surgical videos using only weak surgical phase labels. Further, by incorporating effective intermediate scene representation disentanglement steps within the pipeline, our solution outperforms the SOTA on the CATARACTS dataset by 8% accuracy and 10% F1 score in surgical workflow recognition.

Topological accuracy in medical image segmentation is a highly important property for downstream applications such as network analysis and flow modeling in vessels or cell counting. Recently, significant methodological advancements have brought well-founded concepts from algebraic topology to binary segmentation. However, these approaches have been underexplored in multi-class segmentation scenarios, where topological errors are common. We propose a general loss function for topologically faithful multi-class segmentation extending the recent Betti matching concept, which is based on induced matchings of persistence barcodes. We project the N-class segmentation problem to N single-class segmentation tasks, which allows us to use 1-parameter persistent homology, making training of neural networks computationally feasible. We validate our method on a comprehensive set of four medical datasets with highly variant topological characteristics. Our loss formulation significantly enhances topological correctness in cardiac, cell, artery-vein, and Circle of Willis segmentation.

Deep learning (DL) methods typically require large datasets to effectively learn data distributions. However, in the medical field, data is often limited in quantity, and acquiring labeled data can be costly. To mitigate this data scarcity, data augmentation techniques are commonly employed. Among these techniques, generative models play a pivotal role in expanding datasets. However, when it comes to ultrasound (US) imaging, the authenticity of generated data often diminishes due to the oversight of ultrasound physics.
We propose a novel approach to improve the quality of generated US images by introducing a physics-based diffusion model that is specifically designed for this image modality. The proposed model incorporates an US-specific scheduler scheme that mimics the natural behavior of sound wave propagation in ultrasound imaging. Our analysis demonstrates how the proposed method aids in modeling the attenuation dynamics in US imaging. We present both qualitative and quantitative results based on standard generative model metrics, showing that our proposed method results in overall more plausible images.

In this work, we introduce Progressive Growing of Patch Size, a resource-efficient implicit curriculum learning approach for dense prediction tasks. Our curriculum approach is defined by growing the patch size during model training, which gradually increases the task’s difficulty. We integrated our curriculum into the nnU-Net framework and evaluated the methodology on all 10 tasks of the Medical Segmentation Decathlon. With our approach, we are able to substantially reduce runtime, computational costs, and emissions of network training compared to classical constant patch size training. In our experiments, the curriculum approach resulted in improved convergence. We are able to outperform standard nnU-Net training, which is trained with constant patch size, in terms of Dice Score on 7 out of 10 MSD tasks while only spending roughly 50% of the original training runtime. To the best of our knowledge, our Progressive Growing of Patch Size is the first successful employment of a sample-length curriculum in the form of patch size in the field of computer vision.

Positron emission tomography (PET) is a well-established functional imaging technique for diagnosing brain disorders. However, PET’s high costs and radiation exposure limit its widespread use. In contrast, magnetic resonance imaging (MRI) does not have these limitations. Although it also captures neurodegenerative changes, MRI is a less sensitive diagnostic tool than PET. To close this gap, we aim to generate synthetic PET from MRI. Herewith, we introduce PASTA, a novel pathology-aware image translation framework based on conditional diffusion models. Compared to the state-of-the-art methods, PASTA excels in preserving both structural and pathological details in the target modality, which is achieved through its highly interactive dual-arm architecture and multi-modal condition integration. A cycle exchange consistency and volumetric generation strategy elevate PASTA’s capability to produce high-quality 3D PET scans. Our qualitative and quantitative results confirm that the synthesized PET scans from PASTA not only reach the best quantitative scores but also preserve the pathology correctly. For Alzheimer’s classification, the performance of synthesized scans improves over MRI by 4%, almost reaching the performance of actual PET.

Physics-inspired regularization is desired for intra-patient image registration since it can effectively capture the biomechanical characteristics of anatomical structures. However, a major challenge lies in the reliance on physical parameters: Parameter estimations vary widely across the literature, and the physical properties themselves are inherently subject-specific. In this work, we introduce a novel data-driven method that leverages hypernetworks to learn the tissue-dependent elasticity parameters of an elastic regularizer. Notably, our approach facilitates the estimation of patient-specific parameters without the need to retrain the network. We evaluate our method on three publicly available 2D and 3D lung CT and cardiac MR datasets. We find that with our proposed subject-specific tissue-dependent regularization, a higher registration quality is achieved across all datasets compared to using a global regularizer.

Ultrasound imaging is challenging to interpret due to non-uniform intensities, low contrast, and inherent artifacts, necessitating extensive training for non-specialists. Advanced representation with clear tissue structure separation could greatly assist clinicians in mapping underlying anatomy and distinguishing between tissue layers. Decomposing an image into semantically meaningful segments is mainly achieved using supervised segmentation algorithms. Unsupervised methods are beneficial, as acquiring large labeled datasets is difficult and costly, but despite their advantages, they still need to be explored in ultrasound. This paper proposes a novel unsupervised deep learning strategy tailored to ultrasound to obtain easily interpretable tissue separations. We integrate key concepts from unsupervised deep spectral methods, which combine spectral graph theory with deep learning methods. We utilize self-supervised transformer features for spectral clustering to generate meaningful segments based on ultrasound-specific metrics and shape and positional priors, ensuring semantic consistency across the dataset. We evaluate our unsupervised deep learning strategy on three ultrasound datasets, showcasing qualitative results across anatomical contexts without label requirements. We also conduct a comparative analysis against other clustering algorithms to demonstrate superior segmentation performance, boundary preservation, and label consistency.

Nuclei semantic segmentation is a key component for advancing machine learning and deep learning applications in digital pathology. However, most existing segmentation models are trained and tested on high-quality data acquired with expensive equipment, such as whole slide scanners, which are not accessible to most pathologists in developing countries. These pathologists rely on low-resource data acquired with low-precision microscopes, smartphones, or digital cameras, which have different characteristics and challenges than high-resource data. Therefore, there is a gap between the state-of-the-art segmentation models and the real-world needs of low-resource settings. This work aims to bridge this gap by presenting the first fully annotated African multi-organ dataset for histopathology nuclei semantic segmentation acquired with a low-precision microscope. We also evaluate state-of-the-art segmentation models, including spectral feature extraction encoder and vision transformer-based models, and stain normalization techniques for color normalization of Hematoxylin and Eosin-stained histopathology slides. Our results provide important insights for future research on nuclei histopathology segmentation with low-resource data.

Every day, countless surgeries are performed worldwide, each within the distinct settings of operating rooms (ORs) that vary not only in their setups but also in the personnel, tools, and equipment used. This inherent diversity poses a substantial challenge for achieving a holistic understanding of the OR, as it requires models to generalize beyond their initial training datasets. To reduce this gap, we introduce ORacle, an advanced vision-language model designed for holistic OR domain modeling, which incorporates multi-view and temporal capabilities and can leverage external knowledge during inference, enabling it to adapt to previously unseen surgical scenarios. This capability is further enhanced by our novel data augmentation framework, which significantly diversifies the training dataset, ensuring ORacle’s proficiency in applying the provided knowledge effectively. In rigorous testing, in scene graph generation, and downstream tasks on the 4D-OR dataset, ORacle not only demonstrates state-of-the-art performance but does so requiring less data than existing models. Furthermore, its adaptability is displayed through its ability to interpret unseen views, actions, and appearances of tools and equipment. This demonstrates ORacle’s potential to significantly enhance the scalability and affordability of OR domain modeling and opens a pathway for future advancements in surgical data science.

In emergency departments, rural hospitals, or clinics in less developed regions, clinicians often lack fast image analysis by trained radiologists, which can have a detrimental effect on patients’ healthcare. Large Language Models (LLMs) have the potential to alleviate some pressure from these clinicians by providing insights that can help them in their decision-making. While these LLMs achieve high test results on medical exams showcasing their great theoretical medical knowledge, they tend not to follow medical guidelines. In this work, we introduce a new approach for zero-shot guideline-driven decision support. We model a system of multiple LLM agents augmented with a contrastive vision-language model that collaborate to reach a patient diagnosis. After providing the agents with simple diagnostic guidelines, they will synthesize prompts and screen the image for findings following these guidelines. Finally, they provide understandable chain-of-thought reasoning for their diagnosis, which is then self-refined to consider inter-dependencies between diseases. As our method is zero-shot, it is adaptable to settings with rare diseases, where training data is limited, but expert-crafted disease descriptions are available. We evaluate our method on two chest X-ray datasets, CheXpert and ChestX-ray 14 Longtail, showcasing performance improvement over existing zero-shot methods and generalizability to rare diseases.

We present a new model for deformable image registration, which learns in an unsupervised way a data-specific similarity metric. The proposed method consists of two neural networks, one that maps pairs of input images to transformations which align them, and one that provides the similarity metric whose maximisation guides the image alignment. We parametrise the similarity metric as an energy-based model, which is simple to train and allows us to improve the accuracy of image registration compared to other models with learnt similarity metrics by taking advantage of a more general mathematical formulation, as well as larger datasets. We also achieve substantial improvement in the accuracy of inter-patient image registration on MRI scans from the OASIS dataset compared to models that rely on traditional functions.

VoxelMorph, proposed in 2018, utilizes Convolutional Neural Networks (CNNs) to address medical image registration problems. In 2021 TransMorph advanced this approach by replacing CNNs with Attention mechanisms, claiming enhanced performance. More recently, the rise of Mamba with selective state space models has led to MambaMorph, which substituted Attention with Mamba blocks, asserting superior registration. These developments prompt a critical question: does chasing the latest computational trends with “more advanced” computational blocks genuinely enhance registration accuracy, or is it merely hype? Furthermore, the role of classic high-level registration-specific designs, such as coarse-to-fine pyramid mechanism, correlation calculation, and iterative optimization, warrants scrutiny, particularly in differentiating their influence from the aforementioned low-level computational blocks. In this study, we critically examine these questions through a rigorous evaluation in brain MRI registration. We employed modularized components for each block and ensured unbiased comparisons across all methods and designs to disentangle their effects on performance. Our findings indicate that adopting “advanced” computational elements fails to significantly improve registration accuracy. Instead, well-established registration-specific designs offer fair improvements, enhancing results by a marginal 1.5% over the baseline. Our findings emphasize the importance of rigorous, unbiased evaluation and contribution disentanglement of all low- and high-level registration components, rather than simply following the computer vision trends with “more advanced” computational blocks. We advocate for simpler yet effective solutions and novel evaluation metrics that go beyond conventional registration accuracy, warranting further research across various organs and modalities.

In this multi-center study, we proposed a structured reporting (SR) framework for non-small cell lung cancer (NSCLC) and developed a software-assisted tool to automatically translate image-based findings and annotations into TNM classifications. The aim of this study was to validate the software-assisted SR tool for NSCLC, assess its potential clinical impact in a proof-of-concept study, and evaluate current reporting standards in participating institutions.

Diagnosing dementia, particularly for Alzheimer’s Disease (AD) and frontotemporal dementia (FTD), is complex due to overlapping symptoms. While magnetic resonance imaging (MRI) and positron emission tomography (PET) data are critical for the diagnosis, integrating these modalities in deep learning faces challenges, often resulting in suboptimal performance compared to using single modalities. Moreover, the potential of multi-modal approaches in differential diagnosis, which holds significant clinical importance, remains largely unexplored. We propose a novel framework, DiaMond, to address these issues with vision Transformers to effectively integrate MRI and PET. DiaMond is equipped with self-attention and a novel bi-attention mechanism that synergistically combine MRI and PET, alongside a multi-modal normalization to reduce redundant dependency, thereby boosting the performance. DiaMond significantly outperforms existing multi-modal methods across various datasets, achieving a balanced accuracy of 92.4% in AD diagnosis, 65.2% for AD-MCI-CN classification, and 76.5% in differential diagnosis of AD and FTD. We also validated the robustness of DiaMond in a comprehensive ablation study.

Recent advances in generative models for medical imaging have shown promise in representing multiple modalities. However, the variability in modality availability across datasets limits the general applicability of the synthetic data they produce. To address this, we present a novel physics-informed generative model capable of synthesizing a variable number of brain MRI modalities, including those not present in the original dataset. Our approach utilizes latent diffusion models and a two-step generative process: first, unobserved physical tissue property maps are synthesized using a latent diffusion model, and then these maps are combined with a physical signal model to generate the final MRI scan. Our experiments demonstrate the efficacy of this approach in generating unseen MR contrasts and preserving physical plausibility. Furthermore, we validate the distributions of generated tissue properties by comparing them to those measured in real brain tissue.

Report generation models offer fine-grained textual interpretations of medical images like chest X-rays, yet they often lack interactivity (i.e. the ability to steer the generation process through user queries) and localized interpretability (i.e. visually grounding their predictions), which we deem essential for future adoption in clinical practice. While there have been efforts to tackle these issues, they are either limited in their interactivity by not supporting textual queries or fail to also offer localized interpretability. Therefore, we propose a novel multitask architecture and training paradigm integrating textual prompts and bounding boxes for diverse aspects like anatomical regions and pathologies. We call this approach the Chest X-Ray Explainer (ChEX). Evaluations across a heterogeneous set of 9 chest X-ray tasks, including localized image interpretation and report generation, showcase its competitiveness with SOTA models while additional analysis demonstrates ChEX’s interactive capabilities.

We present EchoScene, an interactive and controllable generative model that generates 3D indoor scenes on scene graphs. EchoScene leverages a dual-branch diffusion model that dynamically adapts to scene graphs. Existing methods struggle to handle scene graphs due to varying numbers of nodes, multiple edge combinations, and manipulator-induced node-edge operations. EchoScene overcomes this by associating each node with a denoising process and enables collaborative information exchange, enhancing controllable and consistent generation aware of global constraints. This is achieved through an information echo scheme in both shape and layout branches. At every denoising step, all processes share their denoising data with an information exchange unit that combines these updates using graph convolution. The scheme ensures that the denoising processes are influenced by a holistic understanding of the scene graph, facilitating the generation of globally coherent scenes. The resulting scenes can be manipulated during inference by editing the input scene graph and sampling the noise in the diffusion model. Extensive experiments validate our approach, which maintains scene controllability and surpasses previous methods in generation fidelity. Moreover, the generated scenes are of high quality and thus directly compatible with off-the-shelf texture generation. Our code and models are open-sourced.

In minimally invasive endovascular procedures, contrast-enhanced angiography remains the most robust imaging technique. However, it is at the expense of the patient and clinician’s health due to prolonged radiation exposure. As an alternative, interventional ultrasound has notable benefits such as being radiation-free, fast to deploy, and having a small footprint in the operating room. Yet, ultrasound is hard to interpret, and highly prone to artifacts and noise. Additionally, interventional radiologists must undergo extensive training before they become qualified to diagnose and treat patients effectively, leading to a shortage of staff, and a lack of open-source datasets. In this work, we seek to address both problems by introducing a self-supervised deep learning architecture to segment catheters in longitudinal ultrasound images, without demanding any labeled data. The network architecture builds upon AiAReSeg, a segmentation transformer built with the Attention in Attention mechanism, and is capable of learning feature changes across time and space. To facilitate training, we used synthetic ultrasound data based on physics-driven catheter insertion simulations, and translated the data into a unique CT-Ultrasound common domain, CACTUSS, to improve the segmentation performance. We generated ground truth segmentation masks by computing the optical flow between adjacent frames using FlowNet2, and performed thresholding to obtain a binary map estimate. Finally, we validated our model on a test dataset, consisting of unseen synthetic data and images collected from silicon aorta phantoms, thus demonstrating its potential for applications to clinical data in the future.

Hintergrund: Die medizinische Codierung von radiologischen Befunden ist essenziell für eine gute Qualität der Versorgung und die korrekte Abrechnung, gleichzeitig aber eine aufwändige und fehleranfällige Aufgabe.
Ziel der Arbeit: Bewertung der Anwendbarkeit natürlicher Sprachverarbeitung (Natural Language Processing, NLP) für die ICD-10-Codierung von radiologischen Befunden in deutscher Sprache durch Finetuning geeigneter Sprachmodelle.
Material und Methoden: In dieser retrospektiven Studie wurden alle Magnetresonanztomographie(MRT)-Befunde unseres Instituts zwischen 2010 und 2020 berücksichtigt. Die ICD-10-Codes bei Entlassung wurden den jeweiligen Befunden zugeordnet, um einen Datensatz für eine Multiclass-Klassifizierung zu erstellen. Finetuning von GermanBERT und flanT5 wurde auf dem Gesamtdatensatz (dstotal) mit 1035 verschiedenen ICD-10-Codes und zwei reduzierten Datensätzen mit den 100 (ds100) und 50 (ds50) häufigsten Codes durchgeführt. Die Performance der Modelle wurde mit Top-k-Genauigkeit für k = 1, 3, 5 evaluiert. In einer Ablationsstudie wurden beide Modelle einmal auf den zugehörigen Metadaten und dem Befund allein trainiert.
Ergebnisse: Der Gesamtdatensatz bestand aus 100.672 radiologischen Befunden, die reduzierten Datensätze ds100 aus 68.103 und ds50 aus 52.293 Berichten. Die Modellperformance stieg, wenn mehrere der besten Voraussagen des Modells in Betracht gezogen wurden, die Anzahl der Zielklassen reduziert wurde und die Metadaten mit dem Befund kombiniert wurden. FlanT5 übertraf GermanBERT in allen Datensätzen und Metriken und eignet sich am besten als medizinischer Codierungsassistent, wobei eine Top-3-Genauigkeit von fast 70% im realitätsnahen Datensatz dstotal erreicht wurde.
Schlussfolgerung: Finetuning von Sprachmodellen verspricht eine zuverlässige Vorhersage von ICD-10-Codes deutscher radiologischer MRT-Befunde in unterschiedlichen Szenarien. Als Codierungsassistent kann flanT5 medizinischen Codierern helfen, informierte Entscheidungen zu treffen und potenziell ihre Arbeitsbelastung reduzieren.

Medical domain applications require a detailed understanding of the decision making process, in particular when data-driven modeling via machine learning is involved, and quantifying uncertainty in the process adds trust and interpretability to predictive models. However, current uncertainty measures in medical imaging are mostly monolithic and do not distinguish between different sources and types of uncertainty. In this paper, we advocate the distinction between so-called aleatoric and epistemic uncertainty in the medical domain and illustrate its potential in clinical decision making for the case of PET/CT image classification.

Category-level object pose estimation, aiming to predict the 6D pose and 3D size of objects from known categories, typically struggles with large intra-class shape variation. Existing works utilizing mean shapes often fall short of cap-turing this variation. To address this issue, we present Sec-ondPose, a novel approach integrating object-specific ge-ometric features with semantic category priors from DI-NOv2. Leveraging the advantage of DINOv2 in providing SE(3)-consistent semantic features, we hierarchically extract two types of SE(3)-invariant geometric features to further encapsulate local-to-global object-specific information. These geometric features are then point-aligned with DINOv2 features to establish a consistent object represen-tation under SE(3) transformations, facilitating the map-ping from camera space to the pre-defined canonical space, thus further enhancing pose estimation. Extensive exper-iments on NOCS-REAL275 demonstrate that SecondPose achieves a 12.4% leap forward over the state-of-the-art. Moreover, on a more complex dataset HouseCat6D which provides photometrically challenging objects, SecondPose still surpasses other competitors by a large margin.

This paper introduces a novel top-down representation approach for deformable image registration, which estimates the deformation field by capturing various short-and long-range flow features at different scale levels. As a Hierarchical Vision Transformer (H-ViT), we propose a dual self-attention and cross-attention mechanism that uses high-level features in the deformation field to represent low-level ones, enabling information streams in the deformation field across all voxel patch embeddings irrespective of their spatial proximity. Since high-level features contain abstract flow patterns, such patterns are expected to effectively contribute to the representation of the deformation field in lower scales. When the self-attention module utilizes within-scale short-range patterns for representation, the cross-attention modules dynamically look for the key tokens across different scales to further interact with the local query voxel patches. Our method shows superior accuracy and visual quality over the state-of-the-art registration methods in five publicly available datasets, highlighting a substantial enhancement in the performance of medical imaging registration.

Recent learning methods for object pose estimation require resource-intensive training for each individual object instance or category, hampering their scalability in real applications when confronted with previously unseen objects. In this paper, we propose MatchU, a Fuse-Describe-Match strategy for 6D pose estimation from RGB-D images. MatchU is a generic approach that fuses 2D texture and 3D geometric cues for 6D pose prediction of unseen objects. We rely on learning geometric 3D descriptors that are rotation-invariant by design. By encoding pose-agnostic geometry, the learned descriptors naturally generalize to unseen objects and capture symmetries. To tackle ambiguous associations using 3D geometry only, we fuse additional RGB information into our descriptor. This is achieved through a novel attention-based mechanism that fuses cross-modal information, together with a matching loss that leverages the latent space learned from RGB data to guide the descriptor learning process. Extensive experiments reveal the generalizability of both the RGB-D fusion strategy as well as the descriptor efficacy. Benefiting from the novel designs, MatchU surpasses all existing methods by a significant margin in terms of both accuracy and speed, even without the requirement of expensive re-training or rendering.

Estimating 6D object poses is a major challenge in 3D computer vision. Building on successful instance-level approaches, research is shifting towards category-level pose estimation for practical applications. Current category-level datasets, however, fall short in annotation quality and pose variety. Addressing this, we introduce HouseCat6D, a new category-level 6D pose dataset. It features 1) multi-modality with Polarimetric RGB and Depth (RGBD+P), 2) encompasses 194 diverse objects across 10 household cat-egories, including two photometrically challenging ones, and 3) provides high-quality pose annotations with an error range of only 1.35 mm to 1.74 mm. The dataset also includes 4) 41 large-scale scenes with comprehensive view-point and occlusion coverage,5) a checkerboard-free en-vironment, and 6) dense 6D parallel-jaw robotic grasp annotations. Additionally, we present benchmark results for leading category-level pose estimation networks.

Pathological lymph node delineation is crucial in cancer diagnosis, progression assessment, and treatment planning. The MICCAI 2023 Lymph Node Quantification Challenge published the first public dataset for pathological lymph node segmentation in the mediastinum. As lymph node annotations are expensive, the challenge was formed as a weakly supervised learning task, where only a subset of all lymph nodes in the training set have been annotated. For the challenge submission, multiple methods for training on these weakly supervised data were explored, including noisy label training, loss masking of unlabeled data, and an approach that integrated the TotalSegmentator toolbox as a form of pseudo labeling in order to reduce the number of unknown voxels. Furthermore, multiple public TCIA datasets were incorporated into the training to improve the performance of the deep learning model. Our submitted model achieved a Dice score of 0.628 and an average symmetric surface distance of 5.8~mm on the challenge test set. With our submitted model, we accomplished the third rank in the MICCAI2023 LNQ challenge. A finding of our analysis was that the integration of all visible, including non-pathological lymph nodes improved the overall segmentation performance on pathological lymph nodes of the test set. Furthermore, segmentation models trained only on clinically enlarged lymph nodes, as given in the challenge scenario, could not generalize to smaller pathological lymph nodes.

The tumor grading of patients suffering from soft-tissue sarcomas is a critical task, as an accurate classification of this high-mortality cancer entity constitutes a decisive factor in devising optimal treatment strategies. In this work, we focus on distinguishing soft-tissue sarcoma subtypes solely based on their 3D morphological characteristics, derived from tumor segmentation masks. Notably, we direct attention to overcoming the limitations of texture-based methodologies, which often fall short of providing adequate shape delineation. To this end, we propose a novel yet elegant modular geometric deep learning framework coined Global Local Graph Convolutional Network (GloLo-GCN) that integrates local and global shape characteristics into a meaningful unified shape descriptor. Evaluated on a multi-center dataset, our proposed model performs better in soft-tissue sarcoma grading than GCNs based on state-of-the-art graph convolutions and a volumetric 3D convolutional neural network, also evaluated on binary segmentation masks exclusively.

Cardiac magnetic resonance (CMR) image acquisition requires subjects to hold their breath while 2D cine images are acquired. This process assumes that the heart remains in the same position across all slices. However, differences in breathhold positions or patient motion introduce 3D slice misalignments. In this work, we propose an algorithm that simultaneously aligns all SA and LA slices by maximizing the pair-wise intensity agreement between their intersections. Unlike previous works, our approach is formulated as a subject-specific optimization problem and requires no prior knowledge of the underlying anatomy. We quantitatively demonstrate that the proposed method is robust against a large range of rotations and translations by synthetically misaligning 10 motion-free datasets and aligning them back using the proposed method.

The prevailing deep learning-based methods of predicting cardiac segmentation involve reconstructed magnetic resonance (MR) images. The heavy dependency of segmentation approaches on image quality significantly limits the acceleration rate in fast MR reconstruction. Moreover, the practice of treating reconstruction and segmentation as separate sequential processes leads to artifact generation and information loss in the intermediate stage. These issues pose a great risk to achieving high-quality outcomes. To leverage the redundant k-space information overlooked in this dual-step pipeline, we introduce a novel approach to directly deriving segmentations from sparse k-space samples using a transformer (DiSK). DiSK operates by globally extracting latent features from 2D+time k-space data with attention blocks and subsequently predicting the segmentation label of query points. We evaluate our model under various acceleration factors (ranging from 4 to 64) and compare against two image-based segmentation baselines. Our model consistently outperforms the baselines in Dice and Hausdorff distances across foreground classes for all presented sampling rates.

Ultrasound (US) imaging is widely used in diagnosing and staging abdominal diseases due to its lack of non-ionizing radiation and prevalent availability. However, significant inter-operator variability and inconsistent image acquisition hinder the widespread adoption of extensive screening programs. Robotic ultrasound systems have emerged as a promising solution, offering standardized acquisition protocols and the possibility of automated acquisition. Additionally, these systems enable access to 3D data via robotic tracking, enhancing volumetric reconstruction for improved ultrasound interpretation and precise disease diagnosis.However, the interpretability of 3D US reconstruction of abdominal images can be affected by the patient’s breathing motion. This study introduces a method to compensate for breathing motion in 3D US compounding by leveraging implicit neural representations. Our approach employs a robotic ultrasound system for automated screenings. To demonstrate the method’s effectiveness, we evaluate our proposed method for the diagnosis and monitoring of abdominal aorta aneurysms as a representative use case.Our experiments demonstrate that our proposed pipeline facilitates robust automated robotic acquisition, mitigating artifacts from breathing motion, and yields smoother 3D reconstructions for enhanced screening and medical diagnosis.

The incidence of colorectal cancer (CRC), one of the deadliest cancers around the world, is increasing. Tissue microenvironment (TME) features such as tumor-infiltrating lymphocytes (TILs) can have a crucial impact on diagnosis or decision-making for treating patients with CRC. While clinical studies showed that TILs improve the host immune response, leading to a better prognosis, inter-observer agreement for quantifying TILs is not perfect. Incorporating machine learning (ML) based applications in clinical routine may promote diagnosis reliability. Recently, ML has shown potential for making progress in routine clinical procedures. We aim to systematically review the TILs analysis based on ML in CRC histological images. Deep learning (DL) and non-DL techniques can aid pathologists in identifying TILs, and automated TILs are associated with patient outcomes. However, a large multi-institutional CRC dataset with a diverse and multi-ethnic population is necessary to generalize ML methods.

Objectives: To assess the quality of simplified radiology reports generated with the large language model (LLM) ChatGPT and to discuss challenges and chances of ChatGPT-like LLMs for medical text simplification.
Methods: In this exploratory case study, a radiologist created three fictitious radiology reports which we simplified by prompting ChatGPT with ‘Explain this medical report to a child using simple language.’’ In a questionnaire, we tasked 15 radiologists to rate the quality of the simplified radiology reports with respect to their factual correctness, completeness, and potential harm for patients. We used Likert scale analysis and inductive free-text categorization to assess the quality of the simplified reports.
Results: Most radiologists agreed that the simplified reports were factually correct, complete, and not potentially harmful to the patient. Nevertheless, instances of incorrect statements, missed relevant medical information, and potentially harmful passages were reported.
Conclusion: While we see a need for further adaption to the medical field, the initial insights of this study indicate a tremendous potential in using LLMs like ChatGPT to improve patient-centered care in radiology and other medical domains.
Clinical relevance statement: Patients have started to use ChatGPT to simplify and explain their medical reports, which is expected to affect patient-doctor interaction. This phenomenon raises several opportunities and challenges for clinical routine.

Segmenting ultrasound images is important for precise area and/or volume calculations, ensuring reliable diagnosis and effective treatment evaluation for diseases. Recently, many segmentation methods have been proposed and shown impressive performance. However, currently, there is no deeper understanding of how networks segment target regions or how they define the boundaries. In this paper, we present a new approach that analyzes ultrasound segmentation networks in terms of learned borders because border delimitation is challenging in ultrasound.

Ultrasound (US) imaging, while advantageous for its radiation-free nature, is challenging to interpret due to only partially visible organs and a lack of complete 3D information. While performing US-based diagnosis or investigation, medical professionals therefore create a mental map of the 3D anatomy. In this work, we aim to replicate this process and enhance the visual representation of anatomical structures.

Differential diagnosis of dementia is challenging due to overlapping symptoms, with structural magnetic resonance imaging (MRI) being the primary method for diagnosis. Despite the clinical value of computer-aided differential diagnosis, research has been limited, mainly due to the absence of public datasets that contain diverse types of dementia. This leaves researchers with small in-house datasets that are insufficient for training deep neural networks (DNNs). Self-supervised learning shows promise for utilizing unlabeled MRI scans in training, but small batch sizes for volumetric brain scans make its application challenging. To address these issues, we propose Triplet Training for differential diagnosis with limited target data. It consists of three key stages: (i) self-supervised pre-training on unlabeled data with Barlow Twins, (ii) self-distillation on task-related data, and (iii) fine-tuning on the target dataset. Our approach significantly outperforms traditional training strategies, achieving a balanced accuracy of 75.6%. We further provide insights into the training process by visualizing changes in the latent space after each step. Finally, we validate the robustness of Triplet Training in terms of its individual components in a comprehensive ablation study.

We address the computational barrier of deploying advanced deep learning segmentation models in clinical settings by studying the efficacy of network compression through tensor decomposition. We propose a post-training Tucker factorization that enables the decomposition of pre-existing models to reduce computational requirements without impeding segmentation accuracy. We applied Tucker decomposition to the convolutional kernels of the TotalSegmentator (TS) model, an nnU-Net model trained on a comprehensive dataset for automatic segmentation of 117 anatomical structures. Our approach reduced the floating-point operations (FLOPs) and memory required during inference, offering an adjustable trade-off between computational efficiency and segmentation quality. This study utilized the publicly available TS dataset, employing various downsampling factors to explore the relationship between model size, inference speed, and segmentation performance. The application of Tucker decomposition to the TS model substantially reduced the model parameters and FLOPs across various compression rates, with limited loss in segmentation accuracy. We removed up to 88% of the model’s parameters with no significant performance changes in the majority of classes after fine-tuning. Practical benefits varied across different graphics processing unit (GPU) architectures, with more distinct speed-ups on less powerful hardware. Post-hoc network compression via Tucker decomposition presents a viable strategy for reducing the computational demand of medical image segmentation models without substantially sacrificing accuracy. This approach enables the broader adoption of advanced deep learning technologies in clinical practice, offering a way to navigate the constraints of hardware capabilities.

Magnetic resonance imaging (MRI) is critical for diagnosing neurodegenerative diseases, yet accurately assessing mild cortical atrophy remains a challenge due to its subtlety. Automated cortex reconstruction, paired with healthy reference ranges, aids in pinpointing pathological atrophy, yet their generalization is limited by biases from image acquisition and processing. We introduce the concept of stochastic cortical self-reconstruction (SCSR) that creates a subject-specific healthy reference by taking MRI-derived thicknesses as input and, therefore, implicitly accounting for potential confounders. SCSR randomly corrupts parts of the cortex and self-reconstructs them from the remaining information. Trained exclusively on healthy individuals, repeated self-reconstruction generates a stochastic reference cortex for assessing deviations from the norm. We present three implementations of this concept: XGBoost applied on parcels, and two autoencoders on vertex level – one based on a multilayer perceptron and the other using a spherical U-Net. These models were trained on healthy subjects from the UK Biobank and subsequently evaluated across four public Alzheimer’s datasets. Finally, we deploy the model on clinical in-house data, where deviation maps’ high spatial resolution aids in discriminating between four types of dementia.

Explaining predictions of black-box neural networks is crucial when applied to decision-critical tasks. Thus, attribution maps are commonly used to identify important image regions, despite prior work showing that humans prefer explanations based on similar examples. To this end, ProtoPNet learns a set of class-representative feature vectors (prototypes) for case-based reasoning. During inference, similarities of latent features to prototypes are linearly classified to form predictions and attribution maps are provided to explain the similarity. In this work, we evaluate whether architectures for case-based reasoning fulfill established axioms required for faithful explanations using the example of ProtoPNet. We show that such architectures allow the extraction of faithful explanations. However, we prove that the attribution maps used to explain the similarities violate the axioms. We propose a new procedure to extract explanations for trained ProtoPNets, named ProtoPFaith. Conceptually, these explanations are Shapley values, calculated on the similarity scores of each prototype. They allow to faithfully answer which prototypes are present in an unseen image and quantify each pixel’s contribution to that presence, thereby complying with all axioms. The theoretical violations of ProtoPNet manifest in our experiments on three datasets (CUB-200-2011, Stanford Dogs, RSNA) and five architectures (ConvNet, ResNet, ResNet50, WideResNet50, ResNeXt50). Our experiments show a qualitative difference between the explanations given by ProtoPNet and ProtoPFaith. Additionally, we quantify the explanations with the Area Over the Perturbation Curve, on which ProtoPFaith outperforms ProtoPNet on all experiments by a factor >10^3.

Medical image registration aims to identify the spatial deformation between images of the same anatomical region and is fundamental to image-based diagnostics and therapy. To date, the majority of the deep learning-based registration methods employ regularizers that enforce global spatial smoothness, e.g., the diffusion regularizer. However, such regularizers are not tailored to the data and might not be capable of reflecting the complex underlying deformation. In contrast, physics-inspired regularizers promote physically plausible deformations. One such regularizer is the linear elastic regularizer, which models the deformation of elastic material. These regularizers are driven by parameters that define the material’s physical properties. For biological tissue, a wide range of estimations of such parameters can be found in the literature, and it remains an open challenge to identify suitable parameter values for successful registration. To overcome this problem and to incorporate physical properties into learning-based registration, we propose to use a hypernetwork that learns the effect of the physical parameters of a physics-inspired regularizer on the resulting spatial deformation field. In particular, we adapt the HyperMorph framework to learn the effect of the two elasticity parameters of the linear elastic regularizer. Our approach enables the efficient discovery of suitable, data-specific physical parameters at test time. To the best of our knowledge, we are the first to use a hypernetwork to learn physics-inspired regularization for medical image registration. We evaluate our approach on 3D intrapatient lung CT images. The results show that the linear elastic regularizer can yield comparable results to the diffusion regularizer in unsupervised learning-based registration while predicting deformations with fewer foldings. With our method, the adaptation of the physical parameters to the data can successfully be performed at test time.

Undersampling is a common method in Magnetic Resonance Imaging (MRI) to subsample the number of data points in k-space, reducing acquisition times at the cost of decreased image quality. A popular approach is to employ undersampling patterns following various strategies, e.g., variable density sampling or radial trajectories. In this work, we propose a method that directly learns the under-sampling masks from data points, thereby also providing task- and domain-specific patterns. To solve the resulting discrete optimization problem, we propose a general optimization routine called ProM: A fully probabilistic, differentiable, versatile, and model-free framework for mask optimization that enforces acceleration factors through a convex constraint. Analyzing knee, brain, and cardiac MRI datasets with our method, we discover that different anatomic regions reveal distinct optimal undersampling masks, demonstrating the benefits of using custom masks, tailored for a downstream task. For example, ProM can create undersampling masks that maximize performance in downstream tasks like segmentation with networks trained on fully-sampled MRIs. Even with extreme acceleration factors, ProM yields reasonable performance while being more versatile than existing methods, paving the way for data-driven all-purpose mask generation

Background: Radiological age assessment using reference studies is inherently limited in accuracy due to a finite number of assignable skeletal maturation stages. To overcome this limitation, we present a deep learning approach for continuous age assessment based on clavicle ossification in computed tomography (CT).
Methods: Thoracic CT scans were retrospectively collected from the picture archiving and communication system. Individuals aged 15.0 to 30.0 years examined in routine clinical practice were included. All scans were automatically cropped around the medial clavicular epiphyseal cartilages. A deep learning model was trained to predict a person’s chronological age based on these scans. Performance was evaluated using mean absolute error (MAE). Model performance was compared to an optimistic human reader performance estimate for an established reference study method.
Results: The deep learning model was trained on 4,400 scans of 1,935 patients (training set: mean age =
24.2 years ± 4.0, 1132 female) and evaluated on 300 scans of 300 patients with a balanced age and sex distribution (test set: mean age = 22.5 years ± 4.4, 150 female). Model MAE was 1.65 years, and the highest absolute error was 6.40 years for females and 7.32 years for males. However, performance could be attributed to norm-variants or pathologic disorders. Human reader estimate MAE was 1.84 years and the highest absolute error was 3.40 years for females and 3.78 years for males.
Conclusions: We present a deep learning approach for continuous age predictions using CT volumes highlighting the medial clavicular epiphyseal cartilage with performance comparable to the human reader estimate.

The reconstruction of cortical surfaces is a prerequisite for quantitative analyses of the cerebral cortex in magnetic resonance imaging (MRI). Existing segmentation-based methods separate the surface registration from the surface extraction, which is computationally inefficient and prone to distortions. We introduce Vox2Cortex-Flow (V2C-Flow), a deep mesh-deformation technique that learns a deformation field from a brain template to the cortical surfaces of an MRI scan. To this end, we present a geometric neural network that models the deformation-describing ordinary differential equation in a continuous manner. The network architecture comprises convolutional and graph-convolutional layers, which allows it to work with images and meshes at the same time. V2C-Flow is not only very fast, requiring less than two seconds to infer all four cortical surfaces, but also establishes vertex-wise correspondences to the template during reconstruction. In addition, V2C-Flow is the first approach for cortex reconstruction that models white matter and pial surfaces jointly, therefore avoiding intersections between them. Our comprehensive experiments on internal and external test data demonstrate that V2C-Flow results in cortical surfaces that are state-of-the-art in terms of accuracy. Moreover, we show that the established correspondences are more consistent than in FreeSurfer and that they can directly be utilized for cortex parcellation and group analyses of cortical thickness.

Reconstructing the cortex from longitudinal MRI is indispensable for analyzing morphological changes in the human brain. Despite the recent disruption of cortical surface reconstruction with deep learning, challenges arising from longitudinal data are still persistent. Especially the lack of strong spatiotemporal point correspondence hinders downstream analyses due to the introduced noise. To address this issue, we present V2C-Long, the first dedicated deep learning-based cortex reconstruction method for longitudinal MRI. In contrast to existing methods, V2C-Long surfaces are directly comparable in a cross-sectional and longitudinal manner. We establish strong inherent spatiotemporal correspondences via a novel composition of two deep mesh deformation networks and fast aggregation of feature-enhanced within-subject templates. The results on internal and external test data demonstrate that V2C-Long yields cortical surfaces with improved accuracy and consistency compared to previous methods. Finally, this improvement manifests in higher sensitivity to regional cortical atrophy in Alzheimer’s disease.

Recent years have witnessed a surge of interest in integrating high-dimensional data captured by multisource sensors, driven by the impressive success of neural networks in integrating multimodal data. However, the integration of heterogeneous multimodal data poses a significant challenge, as confounding effects and dependencies among such heterogeneous data sources introduce unwanted variability and bias, leading to suboptimal performance of multimodal models. Therefore, it becomes crucial to normalize the low- or high-level features extracted from data modalities before their fusion takes place. This paper introduces RegBN, a novel approach for multimodal Batch Normalization with REGularization. RegBN uses the Frobenius norm as a regularizer term to address the side effects of confounders and underlying dependencies among different data sources. The proposed method generalizes well across multiple modalities and eliminates the need for learnable parameters, simplifying training and inference. We validate the effectiveness of RegBN on eight databases from five research areas, encompassing diverse modalities such as language, audio, image, video, depth, tabular, and 3D MRI. The proposed method demonstrates broad applicability across different architectures such as multilayer perceptrons, convolutional neural networks, and vision transformers, enabling effective normalization of both low- and high-level features in multimodal neural networks.

Controllable scene synthesis aims to create interactive environments for numerous industrial use cases. Scene graphs provide a highly suitable interface to facilitate these applications by abstracting the scene context in a compact manner. Existing methods, reliant on retrieval from extensive databases or pre-trained shape embeddings, often overlook scene-object and object-object relationships, leading to inconsistent results due to their limited generation capacity. To address this issue, we present CommonScenes, a fully generative model that converts scene graphs into corresponding controllable 3D scenes, which are semantically realistic and conform to commonsense. Our pipeline consists of two branches, one predicting the overall scene layout via a variational auto-encoder and the other generating compatible shapes via latent diffusion, capturing global scene-object and local inter-object relationships in the scene graph while preserving shape diversity. The generated scenes can be manipulated by editing the input scene graph and sampling the noise in the diffusion model. Due to the lack of a scene graph dataset offering high-quality object-level meshes with relations, we also construct SG-FRONT, enriching the off-the-shelf indoor dataset 3D-FRONT with additional scene graph labels. Extensive experiments are conducted on SG-FRONT, where CommonScenes shows clear advantages over other methods regarding generation consistency, quality, and diversity. Codes and the dataset are available on the website.

Aligning 2D ultrasound images with 3D CT scans of the liver holds significant clinical value in enhancing diagnostic precision, surgical planning, and treatment delivery. Conventional approaches primarily rely on optimization techniques, which often have a limited capture range and are susceptible to initialization errors. To address these limitations, we define the problem as “probe pose regression” and leverage deep learning for a more robust and efficient solution for liver US-CT registration without access to paired data. The proposed method is a three-part framework that combines ultrasound rendering, generative model and pose regression. In the first stage, we exploit a differentiable ultrasound rendering model designed to synthesize ultrasound images given segmentation labels. We let the downstream task optimize the rendering parameters, enhancing the performance of the overall method. In the second stage, a generative model bridges the gap between real and rendered ultrasound images, enabling application on real B-mode images. Finally, we use a patient-specific pose regression network, trained self-supervised with only synthetic images and their known poses. We use ultrasound, and CT scans from a dual-modality human abdomen phantom to validate the proposed method.
Our experimental results indicate that the proposed method can estimate probe poses within an acceptable error margin, which can later be fine-tuned using conventional methods. This capability confirms that the proposed framework can serve as a reliable initialization step for US-CT fusion and achieve fully automated US-CT fusion when coupled with conventional methods.

Optimizing a machine learning (ML) pipeline for radiomics analysis involves numerous choices in data set composition, preprocessing, and model selection. Objective identification of the optimal setup is complicated by correlated features, interdependency structures, and a multitude of available ML algorithms. Therefore, we present a radiomics-based benchmarking framework to optimize a comprehensive ML pipeline for the prediction of overall survival. This study is conducted on an image set of patients with hepatic metastases of colorectal cancer, for which radiomics features of the whole liver and of metastases from computed tomography images were calculated. A mixed model approach was used to find the optimal pipeline configuration and to identify the added prognostic value of radiomics features.

Rationale and objectives: We evaluate the automatic identification of type 2 diabetes from neck-to-knee, two-point Dixon MRI scans with 3D convolutional neural networks on a large, population-based dataset. To this end, we assess the best combination of MRI contrasts and stations for diabetes prediction, and the benefit of integrating risk factors.
Materials and methods: Subjects with type 2 diabetes mellitus have been identified in the prospective UK Biobank Imaging study, and a matched control sample has been created to avoid confounding bias. Five-fold cross-validation is used for the evaluation. All scans from the two-point Dixon neck-to-knee sequence have been standardized. A neural network that considers multi-channel MRI input was developed and integrates clinical information in tabular format. An ensemble strategy is used to combine multi-station MRI predictions. A subset with quantitative fat measurements is identified for comparison to prior approaches.
Results: MRI scans from 3406 subjects (mean age, 66.2 years ± 7.1 [standard deviation]; 1128 women) were analyzed with 1703 diabetics. A balanced accuracy of 78.7%, AUC ROC of 0.872, and an average precision of 0.878 was obtained for the classification of diabetes. The ensemble over multiple Dixon MRI stations yields better performance than selecting the individually best station. Moreover, combining fat and water scans as multi-channel inputs to the networks improves upon just using single contrasts as input. Integrating clinical information about known risk factors of diabetes in the network boosts the performance across all stations and the ensemble. The neural network achieved superior results compared to the prediction based on quantitative MRI measurements.
Conclusions: The developed deep learning model accurately predicted type 2 diabetes from neck-to-knee two-point Dixon MRI scans.

Purpose: To analyze and remove protected feature effects in chest radiograph embeddings of deep learning models. Methods: An orthogonalization is utilized to remove the influence of protected features (e.g., age, sex, race) in CXR embeddings, ensuring feature-independent results. To validate the efficacy of the approach, we retrospectively study the MIMIC and CheXpert datasets using three pre-trained models, namely a supervised contrastive, a self-supervised contrastive, and a baseline classifier model. Our statistical analysis involves comparing the original versus the orthogonalized embeddings by estimating protected feature influences and evaluating the ability to predict race, age, or sex using the two types of embeddings. Results: Our experiments reveal a significant influence of protected features on predictions of pathologies. Applying orthogonalization removes these feature effects. Apart from removing any influence on pathology classification, while maintaining competitive predictive performance, orthogonalized embeddings further make it infeasible to directly predict protected attributes and mitigate subgroup disparities. Conclusion: The presented work demonstrates the successful application and evaluation of the orthogonalization technique in the domain of chest X-ray image classification.

Segmentation of anatomical shapes from medical images has taken an important role in the automation of clinical measurements. While typical deep-learning segmentation approaches are performed on discrete voxels, the underlying objects being analysed exist in a real-valued continuous space. Approaches that rely on convolutional neural networks (CNNs) are limited to grid-like inputs and not easily applicable to sparse or partial measurements. We propose a novel family of image segmentation models that tackle many of CNNs’ shortcomings: Neural Implicit Segmentation Functions (NISF). Our framework takes inspiration from the field of neural implicit functions where a network learns a mapping from a real-valued coordinate-space to a shape representation. NISFs have the ability to segment anatomical shapes in high-dimensional continuous spaces. Training is not limited to voxelized grids, and covers applications with sparse and partial data. Interpolation between observations is learnt naturally in the training procedure and requires no post-processing. Furthermore, NISFs allow the leveraging of learnt shape priors to make predictions for regions outside of the original image plane. We go on to show the framework achieves dice scores of on a (3D+t) short-axis cardiac segmentation task using the UK Biobank dataset. We also provide a qualitative analysis on our frameworks ability to perform segmentation and image interpolation on unseen regions of an image volume at arbitrary resolutions.

Inpainting has recently been employed as a successful deep-learning technique for unsupervised model discovery in medical image analysis by taking advantage of the strong priors learned by models to reconstruct the structure and texture of missing parts in images. Even though the learned features depend on the masks as well as the images, the masks used for inpainting are typically random and independent of the dataset, due to the unpredictability of the content of images, i.e., different objects and shapes can appear in different locations in images. However, this is rarely the case for medical imaging data since they are obtained from similar anatomies. Still, random square masks are the most popular technique for inpainting in medical imaging. In this work, we propose a pipeline to generate, position and sample the masks to efficiently learn the shape and structures of the anatomy and generate a myriad of diverse anatomy-aware masks, aiding the model in learning the statistical shape prior to the topology of the organs of interest. We demonstrate the impact of our approach compared to other masking mechanisms in the reconstruction of anatomy. We compare the effectiveness of our proposed masking approach over square-shaped masks, which are traditionally used in medical imaging, and irregular shape masks, which are used in SOTA inpainting literature.

Text-conditioned image generation has made significant progress in recent years with generative adversarial networks and more recently, diffusion models. While diffusion models conditioned on text prompts have produced impressive and high-quality images, accurately representing complex text prompts such as the number of instances of a specific object remains challenging.To address this limitation, we propose a novel guidance approach for the sampling process in the diffusion model that leverages bounding box and segmentation map information at inference time without additional training data. Through a novel loss in the sampling process, our approach guides the model with semantic features from CLIP embeddings and enforces geometric constraints, leading to high-resolution images that accurately represent the scene. To obtain bounding box and segmentation map information, we structure the text prompt as a scene graph and enrich the nodes with CLIP embeddings. Our proposed model achieves state-of-the-art performance on two public benchmarks for image generation from scene graphs, surpassing both scene graph to image and text-based diffusion models in various metrics. Our results demonstrate the effectiveness of incorporating bounding box and segmentation map guidance in the diffusion model sampling process for more accurate text-to-image generation.

Although purely transformer-based architectures pretrained on large datasets are introduced as foundation models for general computer vision tasks, hybrid models that incorporate combinations of convolution and transformer blocks showed state-of-the-art performance in more specialized tasks. Nevertheless, despite the performance gain of both pure and hybrid transformer-based architectures compared to convolutional networks, their high training cost and complexity make it challenging to use them in real scenarios. In this work, we propose a novel and simple architecture based on only convolutional layers and show that by just taking advantage of the attention map visualizations obtained from a self-supervised pretrained vision transformer network, complex transformer-based networks, and even 3D architectures are outperformed with much fewer computation costs. The proposed architecture is composed of two encoder branches with the original image as input in one branch and the attention map visualizations of the same image from multiple self-attention heads from a pre-trained DINO model in the other branch. The results of our experiments on medical imaging datasets show that the extracted attention map visualizations from the attention heads of a pre-trained transformer architecture combined with the image provide strong prior knowledge for a pure CNN architecture to outperform CNN-based and transformer-based architectures.

Abdominal organ segmentation from CT and MRI is an essential prerequisite for surgical planning and computer-aided navigation systems. It is challenging due to the high variability in the shape, size, and position of abdominal organs. Three-dimensional numeric representations of abdominal shapes with point-wise correspondence to a template are further important for quantitative and statistical analyses thereof. Recently, template-based surface extraction methods have shown promising advances for direct mesh reconstruction from volumetric scans. However, the generalization of these deep learning-based approaches to different organs and datasets, a crucial property for deployment in clinical environments, has not yet been assessed. We close this gap and employ template-based mesh reconstruction methods for joint liver, kidney, pancreas, and spleen segmentation. Our experiments on manually annotated CT and MRI data reveal limited generalization capabilities of previous methods to organs of different geometry and weak performance on small datasets. We alleviate these issues with a novel deep diffeomorphic mesh-deformation architecture and an improved training scheme. The resulting method, UNetFlow, generalizes well to all four organs and can be easily fine-tuned on new data. Moreover, we propose a simple registration-based post-processing that aligns voxel and mesh outputs to boost segmentation accuracy.

Radiomics, involving analysis of calculated, quantitative features from medical images with machine learning tools, shares the instability challenge with other high-dimensional data analyses due to variations in the training set. This instability affects model interpretation and feature importance assessment. To enhance stability and interpretability, we introduce grouped feature importance, shedding light on tool limitations and advocating for more reliable radiomics-based analysis methods.

While recent advances in large-scale foundational models show promising results, their application to the medical domain has not yet been explored in detail. In this paper, we progress into the realms of large-scale modeling in medical synthesis by proposing Cheff - a foundational cascaded latent diffusion model, which generates highly-realistic chest radiographs providing state-of-the-art quality on a 1-megapixel scale. We further propose MaCheX, which is a unified interface for public chest datasets and forms the largest open collection of chest X-rays up to date. With Cheff conditioned on radiological reports, we further guide the synthesis process over text prompts and unveil the research area of report-to-chest-X-ray generation.

Although the preservation of shape continuity and physiological anatomy is a natural assumption in the segmentation of medical images, it is often neglected by deep learning methods that mostly aim for the statistical modeling of input data as pixels rather than interconnected structures. In biological structures, however, organs are not separate entities; for example, in reality, a severed vessel is an indication of an underlying problem, but traditional segmentation models are not designed to strictly enforce the continuity of anatomy, potentially leading to inaccurate medical diagnoses. To address this issue, we propose a graph-based approach that enforces the continuity and connectivity of anatomical topology in medical images. Our method encodes the continuity of shapes as a graph constraint, ensuring that the network’s predictions maintain this continuity. We evaluate our method on two public benchmarks on retinal vessel segmentation, showing significant improvements in connectivity metrics compared to traditional methods while getting better or on-par performance on segmentation metrics.

Although purely transformer-based architectures showed promising performance in many computer vision tasks, many hybrid models consisting of CNN and transformer blocks are introduced to fit more specialized tasks. Nevertheless, despite the performance gain of both pure and hybrid transformer-based architectures compared to CNNs in medical imaging segmentation, their high training cost and complexity make it challenging to use them in real scenarios. In this work, we propose simple architectures based on purely convolutional layers, and show that by just taking advantage of the attention map visualizations obtained from a self-supervised pretrained vision transformer network (e.g., DINO) one can outperform complex transformer-based networks with much less computation costs. The proposed architecture is composed of two encoder branches with the original image as input in one branch and the attention map visualizations of the same image from multiple self-attention heads from a pre-trained DINO model (as multiple channels) in the other branch. The results of our experiments on two publicly available medical imaging datasets show that the proposed pipeline outperforms U-Net and the state-of-the-art medical image segmentation models.

Graph representation of objects and their relations in a scene, known as a scene graph, provides a precise and discernible interface to manipulate a scene by modifying the nodes or the edges in the graph. Although existing works have shown promising results in modifying the placement and pose of objects, scene manipulation often leads to losing some visual characteristics like the appearance or identity of objects. In this work, we propose DisPositioNet, a model that learns a disentangled representation for each object for the task of image manipulation using scene graphs in a self-supervised manner. Our framework enables the disentanglement of the variational latent embeddings as well as the feature representation in the graph. In addition to producing more realistic images due to the decomposition of features like pose and identity, our method takes advantage of the probabilistic sampling in the intermediate features to generate more diverse images in object replacement or addition tasks. The results of our experiments show that disentangling the feature representations in the latent manifold of the model outperforms the previous works qualitatively and quantitatively on two public benchmarks.

The lack of sufficient annotated image data is a common issue in medical image segmentation. For some organs and densities, the annotation may be scarce, leading to poor model training convergence, while other organs have plenty of annotated data. In this work, we present MetaMedSeg, a gradient-based meta-learning algorithm that redefines the meta-learning task for the volumetric medical data with the goal of capturing the variety between the slices. We also explore different weighting schemes for gradients aggregation, arguing that different tasks might have different complexity and hence, contribute differently to the initialization. We propose an importance-aware weighting scheme to train our model. In the experiments, we evaluate our method on the medical decathlon dataset by extracting 2D slices from CT and MRI volumes of different organs and performing semantic segmentation. The results show that our proposed volumetric task definition leads to up to improvement in terms of IoU compared to related baselines. The proposed update rule is also shown to improve the performance for complex scenarios where the data distribution of the target organ is very different from the source organs.

Federated learning (FL) is a distributed learning method that offers medical institutes the prospect of collaboration in a global model while preserving the privacy of their patients. Although most medical centers conduct similar medical imaging tasks, their differences, such as specializations, number of patients, and devices, lead to distinctive data distributions. Data heterogeneity poses a challenge for FL and the personalization of the local models. In this work, we investigate an adaptive hierarchical clustering method for FL to produce intermediate semi-global models, so clients with similar data distribution have the chance of forming a more specialized model. Our method forms several clusters consisting of clients with the most similar data distributions; then, each cluster continues to train separately. Inside the cluster, we use meta-learning to improve the personalization of the participants’ models. We compare the clustering approach with classical FedAvg and centralized training by evaluating our proposed methods on the HAM10k dataset for skin lesion classification with extreme heterogeneous data distribution. Our experiments demonstrate significant performance gain in heterogeneous distribution compared to standard FL methods in classification accuracy. Moreover, we show that the models converge faster if applied in clusters and outperform centralized training while using only a small subset of data.

Generative models allow for the creation of highly realistic artificial samples, opening up promising applications in medical imaging. In this work, we propose a multi-stage encoder-based approach to invert the generator of a generative adversarial network (GAN) for high resolution chest radiographs. This gives direct access to its implicitly formed latent space, makes generative models more accessible to researchers, and enables to apply generative techniques to actual patient’s images. We investigate various applications for this embedding, including image compression, disentanglement in the encoded dataset, guided image manipulation, and creation of stylized samples. We find that this type of GAN inversion is a promising research direction in the domain of chest radiograph modeling and opens up new ways to combine realistic X-ray sample synthesis with radiological image analysis.

Automated segmentation of retinal optical coherence tomography (OCT) images has become an important recent direction in machine learning for medical applications. We hypothesize that the anatomic structure of layers and their high-frequency variation in OCT images make retinal OCT a fitting choice for extracting spectral domain features and combining them with spatial domain features. In this work, we present Y-Net, an architecture that combines the frequency domain features with the image domain to improve the segmentation performance of OCT images. The results of this work demonstrate that the introduction of two branches, one for spectral and one for spatial domain features, brings very significant improvement in fluid segmentation performance and allows outperformance as compared to the well-known U-Net model. Our improvement was 13% on the fluid segmentation dice score and 1.9% on the average dice score. Finally, removing selected frequency ranges in the spectral domain demonstrates the impact of these features on the fluid segmentation outperformance.

Do black-box neural network models learn clinically relevant features for fracture diagnosis? The answer not only establishes reliability, quenches scientific curiosity, but also leads to explainable and verbose findings that can assist the radiologists in the final and increase trust. This work identifies the concepts networks use for vertebral fracture diagnosis in CT images. This is achieved by associating concepts to neurons highly correlated with a specific diagnosis in the dataset. The concepts are either associated with neurons by radiologists pre-hoc or are visualized during a specific prediction and left for the user’s interpretation. We evaluate which concepts lead to correct diagnosis and which concepts lead to false positives. The proposed frameworks and analysis pave the way for reliable and explainable vertebral fracture diagnosis.

One challenging property lurking in medical datasets is the imbalanced data distribution, where the frequency of the samples between the different classes is not balanced. Training a model on an imbalanced dataset can introduce unique challenges to the learning problem where a model is biased towards the highly frequent class. Many methods are proposed to tackle the distributional differences and the imbalanced problem. However, the impact of these approaches on the learned features is not well studied. In this paper, we look deeper into the internal units of neural networks to observe how handling data imbalance affects the learned features. We study several popular cost-sensitive approaches for handling data imbalance and analyze the feature maps of the convolutional neural networks from multiple perspectives: analyzing the alignment of salient features with pathologies and analyzing the pathology-related concepts encoded by the networks. Our study reveals differences and insights regarding the trained models that are not reflected by quantitative metrics such as AUROC and AP and show up only by looking at the models through a lens.

Inpainting has recently been proposed as a successful deep learning technique for unsupervised medical image model discovery. The masks used for inpainting are generally independent of the dataset and are not tailored to perform on different given classes of anatomy. In this work, we introduce a method for generating shape-aware masks for inpainting, which aims at learning the statistical shape prior. We hypothesize that although the variation of masks improves the generalizability of inpainting models, the shape of the masks should follow the topology of the organs of interest. Hence, we propose an unsupervised guided masking approach based on an off-the-shelf inpainting model and a superpixel over-segmentation algorithm to generate a wide range of shape-dependent masks. Experimental results on abdominal MR image reconstruction show the superiority of our proposed masking method over standard methods using square-shaped or dataset of irregular shape masks.

It is a mystery which input features contribute to a neural network’s output. Various explanation (feature attribution) methods are proposed in the literature to shed light on the problem. One peculiar observation is that these explanations (attributions) point to different features as being important. The phenomenon raises the question, which explanation to trust? We propose a framework for evaluating the explanations using the neural network model itself. The framework leverages the network to generate input features that impose a particular behavior on the output. Using the generated features, we devise controlled experimental setups to evaluate whether an explanation method conforms to an axiom. Thus we propose an empirical framework for axiomatic evaluation of explanation methods. We evaluate well-known and promising explanation solutions using the proposed framework. The framework provides a toolset to reveal properties and drawbacks within existing and future explanation solutions

The automation of chest X-ray reporting has garnered significant interest due to the time-consuming nature of the task. However, the clinical accuracy of free-text reports has proven challenging to quantify using natural language processing metrics, given the complexity of medical information, the variety of writing styles, and the potential for typos and inconsistencies. Structured reporting and standardized reports, on the other hand, can provide consistency and formalize the evaluation of clinical correctness. However, high-quality annotations for structured reporting are scarce. Therefore, we propose a method to predict clinical findings defined by sentences in structured reporting templates, which can be used to fill such templates. The approach involves training a contrastive language-image model using chest X-rays and related free-text radiological reports, then creating textual prompts for each structured finding and optimizing a classifier to predict clinical findings in the medical image. Results show that even with limited image-level annotations for training, the method can accomplish the structured reporting tasks of severity assessment of cardiomegaly and localizing pathologies in chest X-rays.

One principal approach for illuminating a black-box neural network is feature attribution, i.e. identifying the importance of input features for the network’s prediction. The predictive information of features is recently proposed as a proxy for the measure of their importance. So far, the predictive information is only identified for latent features by placing an information bottleneck within the network. We propose a method to identify features with predictive information in the input domain. The method results in fine-grained identification of input features’ information and is agnostic to network architecture. The core idea of our method is leveraging a bottleneck on the input that only lets input features associated with predictive latent features pass through. We compare our method with several feature attribution methods using mainstream feature attribution evaluation experiments. The code is publicly available.

Deep learning excels in the analysis of unstructured data and recent advancements allow to extend these techniques to survival analysis. In the context of clinical radiology, this enables, e.g., to relate unstructured volumetric images to a risk score or a prognosis of life expectancy and support clinical decision making. Medical applications are, however, associated with high criticality and consequently, neither medical personnel nor patients do usually accept black box models as reason or basis for decisions. Apart from averseness to new technologies, this is due to missing interpretability, transparency and accountability of many machine learning methods. We propose a hazard-regularized variational autoencoder that supports straightforward interpretation of deep neural architectures in the context of survival analysis, a field highly relevant in healthcare. We apply the proposed approach to abdominal CT scans of patients with liver tumors and their corresponding survival times.

The application of deep learning in survival analysis (SA) allows utilizing unstructured and high-dimensional data types uncommon in traditional survival methods. This allows to advance methods in fields such as digital health, predictive maintenance, and churn analysis, but often yields less interpretable and intuitively understandable models due to the black-box character of deep learning-based approaches. We close this gap by proposing 1) a multi-task variational autoencoder (VAE) with survival objective, yielding survival-oriented embeddings, and 2) a novel method HazardWalk that allows to model hazard factors in the original data space. HazardWalk transforms the latent distribution of our autoencoder into areas of maximized/minimized hazard and then uses the decoder to project changes to the original domain. Our procedure is evaluated on a simulated dataset as well as on a dataset of CT imaging data of patients with liver metastases.

Generation of images from scene graphs is a promising direction towards explicit scene generation and manipulation. However, the images generated from the scene graphs lack quality, which in part comes due to high difficulty and diversity in the data. We propose MIGS (Meta Image Generation from Scene Graphs), a meta-learning based approach for few-shot image generation from graphs that enables adapting the model to different scenes and increases the image quality by training on diverse sets of tasks. By sampling the data in a task-driven fashion, we train the generator using meta-learning on different sets of tasks that are categorized based on the scene attributes. Our results show that using this meta-learning approach for the generation of images from scene graphs achieves state-of-the-art performance in terms of image quality and capturing the semantic relationships in the scene.

Despite recent advancements in single-domain or single-object image generation, it is still challenging to generate complex scenes containing diverse, multiple objects and their interactions. Scene graphs, composed of nodes as objects and directed-edges as relationships among objects, offer an alternative representation of a scene that is more semantically grounded than images. We hypothesize that a generative model for scene graphs might be able to learn the underlying semantic structure of real-world scenes more effectively than images, and hence, generate realistic novel scenes in the form of scene graphs. In this work, we explore a new task for the unconditional generation of semantic scene graphs. We develop a deep auto-regressive model called SceneGraphGen which can directly learn the probability distribution over labelled and directed graphs using a hierarchical recurrent architecture. The model takes a seed object as input and generates a scene graph in a sequence of steps, each step generating an object node, followed by a sequence of relationship edges connecting to the previous nodes. We show that the scene graphs generated by SceneGraphGen are diverse and follow the semantic patterns of real-world scenes. Additionally, we demonstrate the application of the generated graphs in image synthesis, anomaly detection and scene graph completion.

Convolutional neural networks are showing promise in the automatic diagnosis of thoracic pathologies on chest x-rays. Their black-box nature has sparked many recent works to explain the prediction via input feature attribution methods (aka saliency methods). However, input feature attribution methods merely identify the importance of input regions for the prediction and lack semantic interpretation of model behavior. In this work, we first identify the semantics associated with internal units (feature maps) of the network. We proceed to investigate the following questions; Does a regression model that is only trained with COVID-19 severity scores implicitly learn visual patterns associated with thoracic pathologies? Does a network that is trained on weakly labeled data (e.g. healthy, unhealthy) implicitly learn pathologies? Moreover, we investigate the effect of pretraining and data imbalance on the interpretability of learned features. In addition to the analysis, we propose semantic attribution to semantically explain each prediction. We present our findings using publicly available chest pathologies (CheXpert [5], NIH ChestX-ray8 [25]) and COVID-19 datasets (BrixIA [20], and COVID-19 chest X-ray segmentation dataset [4]).

Neural networks have demonstrated remarkable performance in classification and regression tasks on chest X-rays. In order to establish trust in the clinical routine, the networks’ prediction mechanism needs to be interpretable. One principal approach to interpretation is feature attribution. Feature attribution methods identify the importance of input features for the output prediction. Building on Information Bottleneck Attribution (IBA) method, for each prediction we identify the chest X-ray regions that have high mutual information with the network’s output. Original IBA identifies input regions that have sufficient predictive information. We propose Inverse IBA to identify all informative regions. Thus all predictive cues for pathologies are highlighted on the X-rays, a desirable property for chest X-ray diagnosis. Moreover, we propose Regression IBA for explaining regression models. Using Regression IBA we observe that a model trained on cumulative severity score labels implicitly learns the severity of different X-ray regions. Finally, we propose Multi-layer IBA to generate higher resolution and more detailed attribution/saliency maps. We evaluate our methods using both human-centric (ground-truth-based) interpretability metrics, and human-agnostic feature importance metrics on NIH Chest X-ray8 and BrixIA datasets.

Background: Yttrium-90 radioembolization (RE) plays an important role in the treatment of liver malignancies. Optimal patient selection is crucial for an effective and safe treatment. In this study, we aim to validate the prognostic performance of a previously established random survival forest (RSF) with an external validation cohort from a different national center. Furthermore, we compare outcome prediction models with different established metrics. Methods: A previously established RSF model, trained on a consecutive cohort of 366 patients who had received RE due to primary or secondary liver tumor at a national center (center 1), was used to predict the outcome of an independent consecutive cohort of 202 patients from a different national center (center 2) and vice versa. Prognostic performance was evaluated using the concordance index (C-index) and the integrated Brier score (IBS). The prognostic importance of designated baseline parameters was measured with the minimal depth concept, and the influence on the predicted outcome was analyzed with accumulated local effects plots. RSF values were compared to conventional cox proportional hazards models in terms of C-index and IBS. Results: The established RSF model achieved a C-index of 0.67 for center 2, comparable to the results obtained for center 1, which it was trained on (0.66). The RSF model trained on center 2 achieved a C-index of 0.68 on center 2 data and 0.66 on center 1 data. CPH models showed comparable results on both cohorts, with C-index ranging from 0.68 to 0.72. IBS validation showed more differentiated results depending on which cohort was trained on and which cohort was predicted (range: 0.08 to 0.20). Baseline cholinesterase was the most important variable for survival prediction. Conclusion: The previously developed predictive RSF model was successfully validated with an independent external cohort. C-index and IBS are suitable metrics to compare outcome prediction models, with IBS showing more differentiated results. The findings corroborate that survival after RE is critically determined by functional hepatic reserve and thus baseline liver function should play a key role in patient selection.

Is critical input information encoded in specific sparse pathways within the neural network? In this work, we discuss the problem of identifying these critical pathways and subsequently leverage them for interpreting the network’s response to an input. The pruning objective — selecting the smallest group of neurons for which the response remains equivalent to the original network — has been previously proposed for identifying critical pathways. We demonstrate that sparse pathways derived from pruning do not necessarily encode critical input information. To ensure sparse pathways include critical fragments of the encoded input information, we propose pathway selection via neurons’ contribution to the response. We proceed to explain how critical pathways can reveal critical input features. We prove that pathways selected via neuron contribution are locally linear (in an ℓ 2 -ball), a property that we use for proposing a feature attribution method: ‘pathway gradient’. We validate our interpretation method using mainstream evaluation experiments. The validation of pathway gradient interpretation method further confirms that selected pathways using neuron contributions correspond to critical input features. The code 1 2 is publicly available.

We propose a versatile framework for survival analysis that combines advanced concepts from statistics with deep learning. The presented framework is based on piecewise expo-nential models and thereby supports various survival tasks, such as competing risks and multi-state modeling, and further allows for estimation of time-varying effects and time-varying features. To also include multiple data sources and higher-order interaction effects into the model, we embed the model class in a neural network and thereby enable the si-multaneous estimation of both inherently interpretable structured regression inputs as well as deep neural network components which can potentially process additional unstructured data sources. A proof of concept is provided by using the framework to predict Alzheimer’s disease progression based on tabular and 3D point cloud data and applying it to synthetic data.

Segmentation of Multiple Sclerosis (MS) lesions in longitudinal brain MR scans is performed for monitoring the progression of MS lesions. We hypothesize that the spatio-temporal cues in longitudinal data can aid the segmentation algorithm. Therefore, we propose a multi-task learning approach by defining an auxiliary self-supervised task of deformable registration between two time-points to guide the neural network toward learning from spatio-temporal changes. We show the efficacy of our method on a clinical dataset comprised of 70 patients with one follow-up study for each patient. Our results show that spatio-temporal information in longitudinal data is a beneficial cue for improving segmentation. We improve the result of current state-of-the-art by 2.6% in terms of overall score (p < 0.05).

Federated learning (FL) has been a promising approach in the field of medical imaging in recent years. A critical problem in FL, specifically in medical scenarios is to have a more accurate shared model which is robust to noisy and out-of distribution clients. In this work, we tackle the problem of statistical heterogeneity in data for FL which is highly plausible in medical data where for example the data comes from different sites with different scanner settings. We propose IDA (Inverse Distance Aggregation), a novel adaptive weighting approach for clients based on meta-information which handles unbalanced and non-iid data. We extensively analyze and evaluate our method against the well-known FL approach, Federated Averaging as a baseline.

Learning disentangled representations from unlabelled data is a fundamental challenge in machine learning. Solving it may unlock other problems, such as generalization, interpretability, or fairness. Although remarkably challenging to solve in theory, disentanglement is often achieved in practice through prior matching. Furthermore, recent works have shown that prior matching approaches can be enhanced by leveraging geometrical considerations, e.g., by learning representations that preserve geometric features of the data, such as distances or angles between points. However, matching the prior while preserving geometric features is challenging, as a mapping that fully preserves these features while aligning the data distribution with the prior does not exist in general. To address these challenges, we introduce a novel approach to disentangled representation learning based on quadratic optimal transport. We formulate the problem using Gromov-Monge maps that transport one distribution onto another with minimal distortion of predefined geometric features, preserving them as much as can be achieved. To compute such maps, we propose the Gromov-Monge-Gap (GMG), a regularizer quantifying whether a map moves a reference distribution with minimal geometry distortion. We demonstrate the effectiveness of our approach for disentanglement across four standard benchmarks, outperforming other methods leveraging geometric considerations.

In optimal transport (OT), a Monge map is known as a mapping that transports a source distribution to a target distribution in the most cost-efficient way. Recently, multiple neural estimators for Monge maps have been developed and applied in diverse unpaired domain translation tasks, e.g. in single-cell biology and computer vision. However, the classic OT framework enforces mass conservation, which makes it prone to outliers and limits its applicability in real-world scenarios. The latter can be particularly harmful in OT domain translation tasks, where the relative position of a sample within a distribution is explicitly taken into account. While unbalanced OT tackles this challenge in the discrete setting, its integration into neural Monge map estimators has received limited attention. We propose a theoretically grounded method to incorporate unbalancedness into any Monge map estimator. We improve existing estimators to model cell trajectories over time and to predict cellular responses to perturbations. Moreover, our approach seamlessly integrates with the OT flow matching (OT-FM) framework. While we show that OT-FM performs competitively in image translation, we further improve performance by incorporating unbalancedness (UOT-FM), which better preserves relevant features. We hence establish UOT-FM as a principled method for unpaired image translation.

Functional gene embeddings, numerical vectors capturing gene function, provide a promising way to integrate functional gene information into machine learning models. These embeddings are learnt by applying self-supervised machine-learning algorithms on various data types including quantitative omics measurements, protein–protein interaction networks and literature. However, downstream evaluations comparing alternative data modalities used to construct functional gene embeddings have been lacking. Here we benchmarked functional gene embeddings obtained from various data modalities for predicting disease-gene lists, cancer drivers, phenotype–gene associations and scores from genome-wide association studies. Off-the-shelf predictors trained on precomputed embeddings matched or outperformed dedicated state-of-the-art predictors, demonstrating their high utility. Embeddings based on literature and protein–protein interactions inferred from low-throughput experiments outperformed embeddings derived from genome-wide experimental data (transcriptomics, deletion screens and protein sequence) when predicting curated gene lists. In contrast, they did not perform better when predicting genome-wide association signals and were biased towards highly-studied genes. These results indicate that embeddings derived from literature and low-throughput experiments appear favourable in many existing benchmarks because they are biased towards well-studied genes and should therefore be considered with caution. Altogether, our study and precomputed embeddings will facilitate the development of machine-learning models in genetics and related fields.

Deep neural networks (DNNs) enable learning from various data modalities, such as images or text. This concept has also found its way into statistical modelling through the use of semi-structured regression, a model additively combining structured predictors with unstructured effects from arbitrary data modalities learned through a DNN. This paper introduces a new framework called sparse modality regression (SMR). SMR is a regression model combining different data modalities and uses a group lasso-type regularization approach to perform modality selection by zeroing out potentially uninformative modalities.

We present a framework for smooth optimization of explicitly regularized objectives for (structured) sparsity. These non-smooth and possibly non-convex problems typically rely on solvers tailored to specific models and regularizers. In contrast, our method enables fully differentiable and approximation-free optimization and is thus compatible with the ubiquitous gradient descent paradigm in deep learning. The proposed optimization transfer comprises an overparameterization of selected parameters and a change of penalties. In the overparametrized problem, smooth surrogate regularization induces non-smooth, sparse regularization in the base parametrization. We prove that the surrogate objective is equivalent in the sense that it not only has identical global minima but also matching local minima, thereby avoiding the introduction of spurious solutions. Additionally, our theory establishes results of independent interest regarding matching local minima for arbitrary, potentially unregularized, objectives. We comprehensively review sparsity-inducing parametrizations across different fields that are covered by our general theory, extend their scope, and propose improvements in several aspects. Numerical experiments further demonstrate the correctness and effectiveness of our approach on several sparse learning problems ranging from high-dimensional regression to sparse neural network training.

Background: The rise of large-scale multi-species genome sequencing projects promises to shed new light on how genomes encode gene regulatory instructions. To this end, new algorithms are needed that can leverage conservation to capture regulatory elements while accounting for their evolution.
Results: Here, we introduce species-aware DNA language models, which we trained on more than 800 species spanning over 500 million years of evolution. Investigating their ability to predict masked nucleotides from context, we show that DNA language models distinguish transcription factor and RNA-binding protein motifs from background non-coding sequence. Owing to their flexibility, DNA language models capture conserved regulatory elements over much further evolutionary distances than sequence alignment would allow. Remarkably, DNA language models reconstruct motif instances bound in vivo better than unbound ones and account for the evolution of motif sequences and their positional constraints, showing that these models capture functional high-order sequence and evolutionary context. We further show that species-aware training yields improved sequence representations for endogenous and MPRA-based gene expression prediction, as well as motif discovery.
Conclusions: Collectively, these results demonstrate that species-aware DNA language models are a powerful, flexible, and scalable tool to integrate information from large compendia of highly diverged genomes.

Models of intercellular communication in tissues are based on molecular profiles of dissociated cells, are limited to receptor–ligand signaling and ignore spatial proximity in situ. We present node-centric expression modeling, a method based on graph neural networks that estimates the effects of niche composition on gene expression in an unbiased manner from spatial molecular profiling data. We recover signatures of molecular processes known to underlie cell communication.

Recent advances in single-cell technologies have enabled high-throughput molecular profiling of cells across modalities and locations. Single-cell transcriptomics data can now be complemented by chromatin accessibility, surface protein expression, adaptive immune receptor repertoire profiling and spatial information. The increasing availability of single-cell data across modalities has motivated the development of novel computational methods to help analysts derive biological insights. As the field grows, it becomes increasingly difficult to navigate the vast landscape of tools and analysis steps. Here, we summarize independent benchmarking studies of unimodal and multimodal single-cell analysis across modalities to suggest comprehensive best-practice workflows for the most common analysis steps. Where independent benchmarks are not available, we review and contrast popular methods. Our article serves as an entry point for novices in the field of single-cell (multi-)omic analysis and guides advanced users to the most recent best practices.

In recent years, unsupervised analysis of microbiome data, such as microbial network analysis and clustering, has increased in popularity. Many new statistical and computational methods have been proposed for these tasks. This multiplicity of analysis strategies poses a challenge for researchers, who are often unsure which method(s) to use and might be tempted to try different methods on their dataset to look for the “best” ones. However, if only the best results are selectively reported, this may cause over-optimism: the “best” method is overly fitted to the specific dataset, and the results might be non-replicable on validation data. Such effects will ultimately hinder research progress. Yet so far, these topics have been given little attention in the context of unsupervised microbiome analysis. In our illustrative study, we aim to quantify over-optimism effects in this context. We model the approach of a hypothetical microbiome researcher who undertakes four unsupervised research tasks: clustering of bacterial genera, hub detection in microbial networks, differential microbial network analysis, and clustering of samples. While these tasks are unsupervised, the researcher might still have certain expectations as to what constitutes interesting results. We translate these expectations into concrete evaluation criteria that the hypothetical researcher might want to optimize. We then randomly split an exemplary dataset from the American Gut Project into discovery and validation sets multiple times. For each research task, multiple method combinations (e.g., methods for data normalization, network generation, and/or clustering) are tried on the discovery data, and the combination that yields the best result according to the evaluation criterion is chosen. While the hypothetical researcher might only report this result, we also apply the “best” method combination to the validation dataset. The results are then compared between discovery and validation data. In all four research tasks, there are notable over-optimism effects; the results on the validation data set are worse compared to the discovery data, averaged over multiple random splits into discovery/validation data. Our study thus highlights the importance of validation and replication in microbiome analysis to obtain reliable results and demonstrates that the issue of over-optimism goes beyond the context of statistical testing and fishing for significance.

Background: Codon optimality is a major determinant of mRNA translation and degradation rates. However, whether and through which mechanisms its effects are regulated remains poorly understood.
Results: Here we show that codon optimality associates with up to 2-fold change in mRNA stability variations between human tissues, and that its effect is attenuated in tissues with high energy metabolism and amplifies with age. Biochemical modeling and perturbation data through oxygen deprivation and ATP synthesis inhibition reveal that cellular energy variations non-uniformly affect the decoding kinetics of different codons.
Conclusions: This new mechanism of codon effect regulation, independent of tRNA regulation, provides a fundamental mechanistic link between cellular energy metabolism and eukaryotic gene expression.

Single-cell transcriptomics enabled the study of cellular heterogeneity in response to perturbations at the resolution of individual cells. However, scaling high-throughput screens (HTSs) to measure cellular responses for many drugs remains a challenge due to technical limitations and, more importantly, the cost of such multiplexed experiments. Thus, transferring information from routinely performed bulk RNA HTS is required to enrich single-cell data meaningfully.We introduce chemCPA, a new encoder-decoder architecture to study the perturbational effects of unseen drugs. We combine the model with an architecture surgery for transfer learning and demonstrate how training on existing bulk RNA HTS datasets can improve generalisation performance. Better generalisation reduces the need for extensive and costly screens at single-cell resolution. We envision that our proposed method will facilitate more efficient experiment designs through its ability to generate in-silico hypotheses, ultimately accelerating drug discovery.

Recommender Systems (RS) pervade many aspects of our everyday digital life. Proposed to work at scale, state-of-the-art RS allow the modeling of thousands of interactions and facilitate highly individualized recommendations. Conceptually, many RS can be viewed as instances of statistical regression models that incorporate complex feature effects and potentially non-Gaussian outcomes. Such structured regression models, including time-aware varying coefficients models, are, however, limited in their applicability to categorical effects and inclusion of a large number of interactions. Here, we propose Factorized Structured Regression (FaStR) for scalable varying coefficient models. FaStR overcomes limitations of general regression models for large-scale data by combining structured additive regression and factorization approaches in a neural network-based model implementation. This fusion provides a scalable framework for the estimation of statistical models in previously infeasible data settings. Empirical results confirm that the estimation of varying coefficients of our approach is on par with state-of-the-art regression techniques, while scaling notably better and also being competitive with other time-aware RS in terms of prediction performance. We illustrate FaStR’s performance and interpretability on a large-scale behavioral study with smartphone user data.

Large single-cell atlases are now routinely generated to serve as references for analysis of smaller-scale studies. Yet learning from reference data is complicated by batch effects between datasets, limited availability of computational resources and sharing restrictions on raw data. Here we introduce a deep learning strategy for mapping query datasets on top of a reference called single-cell architectural surgery (scArches). scArches uses transfer learning and parameter optimization to enable efficient, decentralized, iterative reference building and contextualization of new datasets with existing references without sharing raw data. Using examples from mouse brain, pancreas, immune and whole-organism atlases, we show that scArches preserves biological state information while removing batch effects, despite using four orders of magnitude fewer parameters than de novo integration. scArches generalizes to multimodal reference mapping, allowing imputation of missing modalities. Finally, scArches retains coronavirus disease 2019 (COVID-19) disease variation when mapping to a healthy reference, enabling the discovery of disease-specific cell states. scArches will facilitate collaborative projects by enabling iterative construction, updating, sharing and efficient use of reference atlases.

Computational trajectory inference enables the reconstruction of cell state dynamics from single-cell RNA sequencing experiments. However, trajectory inference requires that the direction of a biological process is known, largely limiting its application to differentiating systems in normal development. Here, we present CellRank (https://cellrank.org) for single-cell fate mapping in diverse scenarios, including regeneration, reprogramming and disease, for which direction is unknown. Our approach combines the robustness of trajectory inference with directional information from RNA velocity, taking into account the gradual and stochastic nature of cellular fate decisions, as well as uncertainty in velocity vectors. On pancreas development data, CellRank automatically detects initial, intermediate and terminal populations, predicts fate potentials and visualizes continuous gene expression trends along individual lineages. Applied to lineage-traced cellular reprogramming data, predicted fate probabilities correctly recover reprogramming outcomes. CellRank also predicts a new dedifferentiation trajectory during postinjury lung regeneration, including previously unknown intermediate cell states, which we confirm experimentally.

Tissue niches are sources of cellular variation and key to understanding both single-cell and tissue phenotypes. The interaction of a cell with its niche can be described through cell communication events. These events cannot be directly observed in molecular profiling assays of single cells and have to be inferred. However, computational models of cell communication and variance attribution defined on data from dissociated tissues suffer from multiple limitations with respect to their ability to define and to identify communication events. We address these limitations using spatial molecular profiling data with node-centric expression modeling (NCEM), a computational method based on graph neural networks which reconciles variance attribution and communication modeling in a single model of tissue niches. We use these models in varying complexity across spatial assays, such as immunohistochemistry and MERFISH, and biological systems to demonstrate that the statistical cell–cell dependencies discovered by NCEM are plausible signatures of known molecular processes underlying cell communication. We identify principles of tissue organisation as cell communication events across multiple datasets using interpretation mechanisms. In the primary motor cortex, we found gene expression variation that is due to niche composition variation across cortical depth. Using the same approach, we also identified niche-dependent cell state variation in CD8 T cells from inflamed colon and colorectal cancer. Finally, we show that NCEMs can be extended to mixed models of explicit cell communication events and latent intrinsic sources of variation in conditional variational autoencoders to yield holistic models of cellular variation in spatial molecular profiling data. Altogether, this graphical model of cellular niches is a step towards understanding emergent tissue phenotypes.

Recent advances in multiplexed single-cell transcriptomics experiments are facilitating the high-throughput study of drug and genetic perturbations. However, an exhaustive exploration of the combinatorial perturbation space is experimentally unfeasible, so computational methods are needed to predict, interpret, and prioritize perturbations. Here, we present the compositional perturbation autoencoder (CPA), which combines the interpretability of linear models with the flexibility of deep-learning approaches for single-cell response modeling. CPA encodes and learns transcriptional drug responses across different cell type, dose, and drug combinations. The model produces easy-to-interpret embeddings for drugs and cell types, which enables drug similarity analysis and predictions for unseen dosage and drug combinations. We show that CPA accurately models single-cell perturbations across compounds, doses, species, and time. We further demonstrate that CPA predicts combinatorial genetic interactions of several types, implying that it captures features that distinguish different interaction programs. Finally, we demonstrate that CPA can generate in-silico 5,329 missing genetic combination perturbations (97.6% of all possibilities) with diverse genetic interactions. We envision our model will facilitate efficient experimental design and hypothesis generation by enabling in-silico response prediction at the single-cell level, and thus accelerate therapeutic applications using single-cell technologies.

Large single-cell atlases are now routinely generated with the aim of serving as reference to analyse future smaller-scale studies. Yet, learning from reference data is complicated by batch effects between datasets, limited availability of computational resources, and sharing restrictions on raw data. Leveraging advances in machine learning, we propose a deep learning strategy to map query datasets on top of a reference called single-cell architectural surgery (scArches, https://github.com/theislab/scarches). It uses transfer learning and parameter optimization to enable efficient, decentralized, iterative reference building, and the contextualization of new datasets with existing references without sharing raw data. Using examples from mouse brain, pancreas, and whole organism atlases, we showcase that scArches preserves nuanced biological state information while removing batch effects in the data, despite using four orders of magnitude fewer parameters compared to de novo integration. To demonstrate mapping disease variation, we show that scArches preserves detailed COVID-19 disease variation upon reference mapping, enabling discovery of new cell identities that are unseen during training. We envision our method to facilitate collaborative projects by enabling the iterative construction, updating, sharing, and efficient use of reference atlases.

Earth Observation (EO) data analysis has been significantly revolutionized by deep learning (DL), with applications typically limited to grid-like data structures. Graph Neural Networks (GNNs) emerge as an important innovation, propelling DL into the non-Euclidean domain. Naturally, GNNs can effectively tackle the challenges posed by diverse modalities, multiple sensors, and the heterogeneous nature of EO data. To introduce GNNs in the related domains, our review begins by offering fundamental knowledge on GNNs. Then, we summarize the generic problems in EO, to which GNNs can offer potential solutions. Following this, we explore a broad spectrum of GNNs’ applications to scientific problems in Earth systems, covering areas such as weather and climate analysis, disaster management, air quality monitoring, agriculture, land cover classification, hydrological process modeling, and urban modeling. The rationale behind adopting GNNs in these fields is explained, alongside methodologies for organizing graphs and designing favorable architectures for various tasks. Furthermore, we highlight methodological challenges of implementing GNNs in these domains and possible solutions that could guide future research. While acknowledging that GNNs are not a universal solution, we conclude the paper by comparing them with other popular architectures like transformers and analyzing their potential synergies.

The investigation of magnetic energy landscapes and the search for ground states of magnetic materials using ab initio methods like density functional theory (DFT) is a challenging task. Complex interactions, such as superexchange and spin-orbit coupling, make these calculations computationally expensive and often lead to non-trivial energy landscapes. Consequently, a comprehensive and systematic investigation of large magnetic configuration spaces is often impractical. We approach this problem by utilizing Bayesian Optimization, an active machine learning scheme that has proven to be efficient in modeling unknown functions and finding global minima. Using this approach we can obtain the magnetic contribution to the energy as a function of one or more spin canting angles with relatively small numbers of DFT calculations. To assess the capabilities and the efficiency of the approach we investigate the noncollinear magnetic energy landscapes of selected materials containing 3d, 5d and 5f magnetic ions: Ba3MnNb2O9, LaMn2Si2, β-MnO2, Sr2IrO4, UO2 and Ba2NaOsO6. By comparing our results to previous ab initio studies that followed more conventional approaches, we observe significant improvements in efficiency.

Making inferences about physical properties of the Universe requires knowledge of the data likelihood. A Gaussian distribution is commonly assumed for the uncertainties with a covariance matrix estimated from a set of simulations. The noise in such covariance estimates causes two problems: it distorts the width of the parameter contours, and it adds scatter to the location of those contours which is not captured by the widths themselves. For non-Gaussian likelihoods, an approximation may be derived via Simulation-Based Inference (SBI). It is often implicitly assumed that parameter constraints from SBI analyses, which do not use covariance matrices, are not affected by the same problems as parameter estimation with a covariance matrix estimated from simulations. We investigate whether SBI suffers from effects similar to those of covariance estimation in Gaussian likelihoods. We use Neural Posterior and Likelihood Estimation with continuous and masked autoregressive normalizing flows for density estimation. We fit our approximate posterior models to simulations drawn from a Gaussian linear model, so that the SBI result can be compared to the true posterior. We test linear and neural network based compression, demonstrating that neither methods circumvent the issues of covariance estimation. SBI suffers an inflation of posterior variance that is equal or greater than the analytical result in covariance estimation for Gaussian likelihoods for the same number of simulations. The assumption that SBI requires a smaller number of simulations than covariance estimation for a Gaussian likelihood analysis is inaccurate. The limitations of traditional likelihood analysis with simulation-based covariance remain for SBI with a finite simulation budget. Despite these issues, we show that SBI correctly draws the true posterior contour given enough simulations.

Transforming CO2 into methanol represents a crucial step towards closing the carbon cycle, with thermoreduction technology nearing industrial application. However, obtaining high methanol yields and ensuring the stability of heterocatalysts remain significant challenges. Herein, we present a sophisticated computational framework to accelerate the discovery of novel thermal heterogeneous catalysts, using machine-learned force fields. We propose a new catalytic descriptor, termed adsorption energy distribution, that aggregates the binding energies for different catalyst facets, binding sites, and adsorbates. The descriptor is versatile and can easily be adjusted to a specific reaction through careful choice of the key-step reactants and reaction intermediates. By applying unsupervised machine learning and statistical analysis to a dataset comprising nearly 160 metallic alloys, we offer a powerful tool for catalyst discovery. Finally, we propose new promising candidate materials such as ZnRh and ZnPt3, which to our knowledge, have not yet been tested, and discuss their possible advantage in terms of stability.

Uncertainty quantification (UQ) is essential for assessing the reliability of Earth observation (EO) products. However, the extensive use of machine learning models in EO introduces an additional layer of complexity, as those models themselves are inherently uncertain. While various UQ methods do exist for machine learning models, their performance on EO datasets remains largely unevaluated. A key challenge in the community is the absence of the ground truth for uncertainty, i.e. how certain the uncertainty estimates are, apart from the labels for the image/signal. This article fills this gap by introducing three benchmark datasets specifically designed for UQ in EO machine learning models. These datasets address three common problem types in EO: regression, image segmentation, and scene classification. They enable a transparent comparison of different UQ methods for EO machine learning models. We describe the creation and characteristics of each dataset, including data sources, preprocessing steps, and label generation, with a particular focus on calculating the reference uncertainty. We also showcase baseline performance of several machine learning models on each dataset, highlighting the utility of these benchmarks for model development and comparison. Overall, this article offers a valuable resource for researchers and practitioners working in artificial intelligence for EO, promoting a more accurate and reliable quality measure of the outputs of machine learning models.

The rapid evolution of Vision Language Models (VLMs) has catalyzed significant advancements in artificial intelligence, expanding research across various disciplines, including Earth Observation (EO). While VLMs have enhanced image understanding and data processing within EO, their applications have predominantly focused on image content description. This limited focus overlooks their potential in geographic and scientific regression tasks, which are essential for diverse EO applications. To bridge this gap, this paper introduces a novel benchmark dataset, called textbf{REO-Instruct} to unify regression and generation tasks specifically for the EO domain. Comprising 1.6 million multimodal EO imagery and language pairs, this dataset is designed to support both biomass regression and image content interpretation tasks. Leveraging this dataset, we develop REO-VLM, a groundbreaking model that seamlessly integrates regression capabilities with traditional generative functions. By utilizing language-driven reasoning to incorporate scientific domain knowledge, REO-VLM goes beyond solely relying on EO imagery, enabling comprehensive interpretation of complex scientific attributes from EO data. This approach establishes new performance benchmarks and significantly enhances the capabilities of environmental monitoring and resource management.

Self-supervised learning guided by masked image modeling, such as masked autoencoder (MAE), has attracted wide attention for pretraining vision transformers in remote sensing. However, MAE tends to excessively focus on pixel details, limiting the model’s capacity for semantic understanding, particularly for noisy synthetic aperture radar (SAR) images. In this article, we explore spectral and spatial remote sensing image features as improved MAE-reconstruction targets. We first conduct a study on reconstructing various image features, all performing comparably well or better than raw pixels. Based on such observations, we propose feature guided MAE (FG-MAE): reconstructing a combination of histograms of oriented gradients (HOG) and normalized difference indices (NDI) for multispectral images, and reconstructing HOG for SAR images. Experimental results on three downstream tasks illustrate the effectiveness of FG-MAE with a particular boost for SAR imagery (e.g., up to 5% better than MAE on EuroSAT-SAR). Furthermore, we demonstrate the well-inherited scalability of FG-MAE and release a first series of pretrained vision transformers for medium-resolution SAR and multispectral images.

Automatically and rapidly understanding Earth’s surface is fundamental to our grasp of the living environment and informed decision-making. This underscores the need for a unified system with comprehensive capabilities in analyzing Earth’s surface to address a wide range of human needs. The emergence of multimodal large language models (MLLMs) has great potential in boosting the efficiency and convenience of intelligent Earth observation. These models can engage in human-like conversations, serve as unified platforms for understanding images, follow diverse instructions, and provide insightful feedbacks. In this study, we introduce LHRS-Bot-Nova, an MLLM specialized in understanding remote sensing (RS) images, designed to expertly perform a wide range of RS understanding tasks aligned with human instructions. LHRS-Bot-Nova features an enhanced vision encoder and a novel bridge layer, enabling efficient visual compression and better language-vision alignment. To further enhance RS-oriented vision-language alignment, we propose a large-scale RS image-caption dataset, generated through feature-guided image recaptioning. Additionally, we introduce an instruction dataset specifically designed to improve spatial recognition abilities. Extensive experiments demonstrate superior performance of LHRS-Bot-Nova across various RS image understanding tasks. We also evaluate different MLLM performances in complex RS perception and instruction following using a complicated multi-choice question evaluation benchmark, providing a reliable guide for future model selection and improvement.

Self-supervised pretraining on large-scale satellite data has raised great interest in building Earth observation (EO) foundation models. However, many important resources beyond pure satellite imagery, such as land-cover-land-use products that provide free global semantic information, as well as vision foundation models that hold strong knowledge of the natural world, are not widely studied. In this work, we show these free additional resources not only help resolve common contrastive learning bottlenecks but also significantly boost the efficiency and effectiveness of EO pretraining. Specifically, we first propose soft contrastive learning (SoftCon) that optimizes cross-scene soft similarity based on land-cover-generated multilabel supervision, naturally solving the issue of multiple positive samples and too strict positive matching in complex scenes. Second, we revisit and explore cross-domain continual pretraining for both multispectral and synthetic aperture radar (SAR) imagery, building efficient EO foundation models from strongest vision models such as DINOv2. Adapting simple weight-initialization and Siamese masking strategies into our SoftCon framework, we demonstrate impressive continual pretraining performance even when the input modalities are not aligned. Without prohibitive training, we produce multispectral and SAR foundation models that achieve significantly better results in 10 out of 11 downstream tasks than most existing SOTA models. For example, our ResNet50/ViT-S achieve 84.8/85.0 linear probing mAP scores on BigEarthNet-10%, which are better than most existing ViT-L models; under the same setting, our ViT-B sets a new record of 86.8 in multispectral, and 82.5 in SAR, the latter even better than many multispectral models.

Domain Generalization (DG) focuses on enhancing the generalization of deep learning models trained on multiple source domains to adapt to unseen target domains. This paper explores DG through the lens of bias-variance decomposition, uncovering that test errors in DG predominantly arise from cross-domain bias and variance. Inspired by this insight, we introduce a Representation Enhancement-Stabilization (RES) framework, comprising a Representation Enhancement (RE) module and a Representation Stabilization (RS) module. In RE, a novel set of feature frequency augmentation techniques is used to progressively reduce cross-domain bias during feature extraction. Furthermore, in RS, a novel Mutual Exponential Moving Average (MEMA) strategy is designed to stabilize model optimization for diminishing cross-domain variance during training. Collectively, the whole RES method can significantly enhance model generalization. We evaluate RES on five benchmark datasets and the results show that it outperforms multiple advanced DG methods.

The increasing availability of multi-sensor data sparks wide interest in multimodal self-supervised learning. However, most existing approaches learn only common representations across modalities while ignoring intra-modal training and modality-unique representations. We propose Decoupling Common and Unique Representations (DeCUR), a simple yet effective method for multimodal self-supervised learning. By distinguishing inter- and intra-modal embeddings through multimodal redundancy reduction, DeCUR can integrate complementary information across different modalities. We evaluate DeCUR in three common multimodal scenarios (radar-optical, RGB-elevation, and RGB-depth), and demonstrate its consistent improvement regardless of architectures and for both multimodal and modality-missing settings. With thorough experiments and comprehensive analysis, we hope this work can provide valuable insights and raise more interest in researching the hidden relationships of multimodal representations.

Exploiting machine learning techniques to automatically classify multispectral remote sensing imagery plays a significant role in deriving changes on the Earth’s surface. However, the computation power required to manage large Earth observation data and apply sophisticated machine learning models for this analysis purpose has become an intractable bottleneck. Leveraging quantum computing provides a possibility to tackle this challenge in the future. This article focuses on land cover classification by analyzing Sentinel-2 images with quantum computing. Two hybrid quantum-classical deep learning frameworks are proposed. Both models exploit quantum computing to extract features efficiently from multispectral images and classical computing for final classification. As proof of concept, numerical simulation results on the LCZ42 dataset through the TensorFlow Quantum platform verify our models’ validity. The experiments indicate that our models can extract features more effectively compared with their classical counterparts, specifically, the convolutional neural network (CNN) model. Our models demonstrated improvements, with an average test accuracy increase of 4.5% and 3.3%, respectively, in comparison to the CNN model. In addition, our proposed models exhibit better transferability and robustness than CNN models.

Monocular height estimation (MHE) is key for generating 3-D city models, essential for swift disaster response. Moving beyond the traditional focus on performance enhancement, our study breaks new ground by probing the interpretability of MHE networks. We have pioneeringly discovered that neurons within MHE models demonstrate selectivity for both height and semantic classes. This insight sheds light on the complex inner workings of MHE models and inspires innovative strategies for leveraging elevation data more effectively. Informed by this insight, we propose a pioneering framework that employs MHE as a self-supervised pretraining method for remote sensing (RS) imagery. This approach significantly enhances the performance of semantic segmentation tasks. Furthermore, we develop a disentangled latent transformer (DLT) module that leverages explainable deep representations from pretrained MHE networks for unsupervised semantic segmentation. Our method demonstrates the significant potential of MHE tasks in developing foundation models for sophisticated pixel-level semantic analyses. Additionally, we present a new dataset designed to benchmark the performance of both semantic segmentation and height estimation tasks.

Change detection (CD) from remote sensing (RS) images using deep learning has been widely investigated in the literature. It is typically regarded as a pixelwise labeling task that aims to classify each pixel as changed or unchanged. Although per-pixel classification networks in encoder-decoder structures have shown dominance, they still suffer from imprecise boundaries and incomplete object delineation at various scenes. For high-resolution RS images, partly or totally changed objects are more worthy of attention rather than a single pixel. Therefore, we revisit the CD task from the mask prediction and classification perspective and propose mask classification-based CD (MaskCD) to detect changed areas by adaptively generating categorized masks from input image pairs. Specifically, it utilizes a cross-level change representation perceiver (CLCRP) to learn multiscale change-aware representations and capture spatiotemporal relations from encoded features by exploiting deformable multihead self-attention (DeformMHSA). Subsequently, a masked cross-attention-based detection transformers (MCA-DETRs) decoder is developed to accurately locate and identify changed objects based on masked cross-attention and self-attention (SA) mechanisms. It reconstructs the desired changed objects by decoding the pixelwise representations into learnable mask proposals and making final predictions from these candidates. Experimental results on five benchmark datasets demonstrate the proposed approach outperforms other state-of-the-art models.

Urban land cover classification aims to derive crucial information from earth observation data and categorize it into specific land uses. To achieve accurate classification, sophisticated machine learning models trained with large earth observation data are employed, but the required computation power has become a bottleneck. Quantum computing might tackle this challenge in the future. However, representing images into quantum states for analysis with quantum computing is challenging due to the high demand for quantum resources. To tackle this challenge, we propose a hybrid quantum neural network that can effectively represent and classify remote sensing imagery with reduced quantum resources. Our model was evaluated on the Local Climate Zone (LCZ)-based land cover classification task using the TensorFlow Quantum platform, and the experimental results indicate its validity for accurate urban land cover classification.

Predicting socioeconomic indicators from satellite imagery with deep learning has become an increasingly popular research direction. Post-hoc concept-based explanations can be an important step towards broader adoption of these models in policy-making as they enable the interpretation of socioeconomic outcomes based on visual concepts that are intuitive to humans. In this paper, we study the interplay between representation learning using an additional task-specific contrastive loss and post-hoc concept explainability for socioeconomic studies. Our results on two different geographical locations and tasks indicate that the task-specific pretraining imposes a continuous ordering of the latent space embeddings according to the socioeconomic outcomes. This improves the model’s interpretability as it enables the latent space of the model to associate urban concepts with continuous intervals of socioeconomic outcomes. Further, we illustrate how analyzing the model’s conceptual sensitivity for the intervals of socioeconomic outcomes can shed light on new insights for urban studies.

Compared to supervised deep learning, self-supervision provides remote sensing a tool to reduce the amount of exact, human-crafted geospatial annotations. While image-level information for unsupervised pretraining efficiently works for various classification downstream tasks, the performance on pixel-level semantic segmentation lags behind in terms of model accuracy. On the contrary, many easily available label sources (e.g., automatic labeling tools and land cover land use products) exist, which can provide a large amount of noisy labels for segmentation model training. In this work, we propose to exploit noisy semantic segmentation maps for model pretraining. Our experiments provide insights on robustness per network layer. The transfer learning settings test the cases when the pretrained encoders are fine-tuned for different label classes and decoders. The results from two datasets indicate the effectiveness of task-specific supervised pretraining with noisy labels. Our findings pave new avenues to improved model accuracy and novel pretraining strategies for efficient remote sensing image segmentation.

The vegetation height has been identified as a key biophysical parameter to justify the role of forests in the carbon cycle and ecosystem productivity. Therefore, consistent and large-scale forest height is essential for managing terrestrial ecosystems, mitigating climate change, and preventing biodiversity loss. Since spaceborne multispectral instruments, Light Detection and Ranging (LiDAR), and Synthetic Aperture Radar (SAR) have been widely used for large-scale earth observation for years, this paper explores the possibility of generating largescale and high-accuracy forest heights with the synergy of the Sentinel-1, Sentinel-2, and ICESat-2 data. A Forest Height Generative Adversarial Network (FH-GAN) is developed to retrieve forest height from Sentinel-1 and Sentinel-2 images sparsely supervised by the ICESat-2 data. This model is made up of a cascade forest height and coherence generator, where the output of the forest height generator is fed into the spatial discriminator to regularize spatial details, and the coherence generator is connected to a coherence discriminator to refine the vertical details. A progressive strategy further underpins the generator to boost the accuracy of multi-source forest height estimation. Results indicated that FH-GAN achieves the best RMSE of 2.10 m at a large scale compared with the LVIS reference and the best RMSE of 6.16 m compared with the ICESat-2 reference.

Urban development in South America has experienced significant growth and transformation over the past few decades. South America’s urban development and trees are closely interconnected, and tree cover within cities plays a vital role in shaping sustainable and resilient urban landscapes. However, knowledge of urban tree canopy (UTC) coverage in the South American continent remains limited. In this study, we used high-resolution satellite images and developed a semi-supervised deep learning method to create UTC data for 888 South American cities. The proposed semi-supervised method can leverage both labeled and unlabeled data during training. By incorporating labeled data for guidance and utilizing unlabeled data to explore underlying patterns, the algorithm enhances model robustness and generalization for urban tree canopy detection across South America, with an average overall accuracy of 94.88% for the tested cities. Based on the created UTC products, we successfully assessed the UTC coverage for each city. Statistical results showed that the UTC coverage in South America is between 0.76% and 69.53%, and the average UTC coverage is approximately 19.99%. Among the 888 cities, only 357 cities that accommodate approximately 48.25% of the total population have UTC coverage greater than 20%, while the remaining 531 cities that accommodate approximately 51.75% of the total population have UTC coverage less than 20%. Natural factors (climatic and geographical) play a very important role in determining UTC coverage, followed by human activity factors (economy and urbanization level). We expect that the findings of this study and the created UTC dataset will help formulate policies and strategies to promote sustainable urban forestry, thus further improving the quality of life of residents in South America.

We study the potential of noisy labels y to pretrain semantic segmentation models in a multi-modal learning framework for geospatial applications. Specifically, we propose a novel Cross-modal Sample Selection method (CromSS) that utilizes the class distributions P^{(d)}(x,c) over pixels x and classes c modelled by multiple sensors/modalities d of a given geospatial scene. Consistency of predictions across sensors d is jointly informed by the entropy of P^{(d)}(x,c). Noisy label sampling we determine by the confidence of each sensor d in the noisy class label, P^{(d)}(x,c=y(x)). To verify the performance of our approach, we conduct experiments with Sentinel-1 (radar) and Sentinel-2 (optical) satellite imagery from the globally-sampled SSL4EO-S12 dataset. We pair those scenes with 9-class noisy labels sourced from the Google Dynamic World project for pretraining. Transfer learning evaluations (downstream task) on the DFC2020 dataset confirm the effectiveness of the proposed method for remote sensing image segmentation.

Uncertainty in machine learning models is a timely and vast field of research. In supervised learning, uncertainty can already occur in the first stage of the training process, the annotation phase. This scenario is particularly evident when some instances cannot be definitively classified. In other words, there is inevitable ambiguity in the annotation step and hence, not necessarily a ‘ground truth’ associated with each instance. The main idea of this work is to drop the assumption of a ground truth label and instead embed the annotations into a multidimensional space. This embedding is derived from the empirical distribution of annotations in a Bayesian setup, modeled via a Dirichlet-Multinomial framework. We estimate the model parameters and posteriors using a stochastic Expectation Maximization algorithm with Markov Chain Monte Carlo steps. The methods developed in this paper readily extend to various situations where multiple annotators independently label instances. To showcase the generality of the proposed approach, we apply our approach to three benchmark datasets for image classification and Natural Language Inference. Besides the embeddings, we can investigate the resulting correlation matrices, which reflect the semantic similarities of the original classes very well for all three exemplary datasets.

Foundation models have enormous potential in advancing Earth and climate sciences, however, current approaches may not be optimal as they focus on a few basic features of a desirable Earth and climate foundation model. Crafting the ideal Earth foundation model, we define eleven features which would allow such a foundation model to be beneficial for any geoscientific downstream application in an environmental- and human-centric this http URL further shed light on the way forward to achieve the ideal model and to evaluate Earth foundation models. What comes after foundation models? Energy efficient adaptation, adversarial defenses, and interpretability are among the emerging directions.

The quantification of predictive uncertainties helps to understand where the existing models struggle to find the correct prediction. A useful quality control tool is the task of detecting out-of-distribution (OOD) data by examining the model’s predictive uncertainty. For this task, deterministic single forward pass frameworks have recently been established as deep learning models and have shown competitive performance in certain tasks. The unique combination of spectrally normalized weight matrices and residual connection networks with an approximate Gaussian process (GP) output layer can here offer the best trade-off between performance and complexity. We utilize this framework with a refined version that adds spectral batch normalization and an inducing points approximation of the GP for the task of OOD detection in remote sensing image classification. This is an important task in the field of remote sensing, because it provides an evaluation of how reliable the model’s predictive uncertainty estimates are. By performing experiments on the benchmark datasets Eurosat and So2Sat LCZ42, we can show the effectiveness of the proposed adaptions to the residual networks (ResNets). Depending on the chosen dataset, the proposed methodology achieves OOD detection performance up to 16% higher than previously considered distance-aware networks. Compared with other uncertainty quantification methodologies, the results are on the same level and exceed them in certain experiments by up to 2%. In particular, spectral batch normalization, which normalizes the batched data as opposed to normalizing the network weights by the spectral normalization (SN), plays a crucial role and leads to performance gains of up to 3% in every single experiment.

The remarkable achievements of ChatGPT and Generative Pre-trained Transformer 4 (GPT-4) have sparked a wave of interest and research in the field of large language models (LLMs) for artificial general intelligence (AGI). These models provide intelligent solutions that are closer to human thinking, enabling us to use general artificial intelligence (AI) to solve problems in various applications. However, in the field of remote sensing (RS), the scientific literature on the implementation of AGI remains relatively scant. Existing AI-related research in RS focuses primarily on visual-understanding tasks while neglecting the semantic understanding of the objects and their relationships. This is where vision-LMs (VLMs) excel as they enable reasoning about images and their associated textual descriptions, allowing for a deeper understanding of the underlying semantics. VLMs can go beyond visual recognition of RS images and can model semantic relationships as well as generate natural language descriptions of the image. This makes them better suited for tasks that require both visual and textual understanding, such as image captioning and visual question answering (VQA). This article provides a comprehensive review of the research on VLMs in RS, summarizing the latest progress, highlighting current challenges, and identifying potential research opportunities. Specifically, we review the application of VLMs in mainstream RS tasks, including image captioning, text-based image generation, text-based image retrieval (TBIR), VQA, scene classification, semantic segmentation, and object detection. For each task, we analyze representative works and discuss research progress. Finally, we summarize the limitations of existing works and provide possible directions for future development. This review aims to provide a comprehensive overview of the current research progress of VLMs in RS (see Figure 1 ), and to inspire further research in this exciting and promising field.

Deep neural networks based on unrolled iterative algorithms have achieved remarkable success in sparse reconstruction applications, such as synthetic aperture radar (SAR) tomographic inversion (TomoSAR). However, the currently available deep learning-based TomoSAR algorithms are limited to 3-D reconstruction. The extension of deep learning-based algorithms to 4-D imaging, i.e., differential TomoSAR (D-TomoSAR) applications, is impeded mainly due to the high-dimensional weight matrices required by the network designed for D-TomoSAR inversion, which typically contain millions of freely trainable parameters. Learning such huge number of weights requires an enormous number of training samples, resulting in a large memory burden and excessive time consumption. To tackle this issue, we propose an efficient and accurate algorithm called HyperLISTA-ABT. The weights in HyperLISTA-ABT are determined in an analytical way according to a minimum coherence criterion, trimming the model down to an ultra-light one with only three hyperparameters. Additionally, HyperLISTA-ABT improves the global thresholding by utilizing an adaptive blockwise thresholding (ABT) scheme, which applies block-coordinate techniques and conducts thresholding in local blocks, so that weak expressions and local features can be retained in the shrinkage step layer by layer. Simulations were performed and demonstrated the effectiveness of our approach, showing that HyperLISTA-ABT achieves superior computational efficiency with no significant performance degradation compared to the state-of-the-art methods. Real data experiments showed that a high-quality 4-D point cloud could be reconstructed over a large area by the proposed HyperLISTA-ABT with affordable computational resources and in a fast time.

Trees in urban areas act as carbon sinks and provide ecosystem services for residents. However, the impact of urbanization on tree coverage in South America remains poorly understood. Here, we make use of very high resolution satellite imagery to derive urban tree coverage for 882 cities in South America and developed a tree coverage impacted (TCI) coefficient to quantify the direct and indirect impacts of urbanization on urban tree canopy (UTC) coverage. The direct effect refers to the change in tree cover due to the rise in urban intensity compared to scenarios with extremely low levels of urbanization, while the indirect impact refers to the change in tree coverage resulting from human management practices and alterations in urban environments. Our study revealed the negative direct impacts and prevalent positive indirect impacts of urbanization on UTC coverage. In South America, 841 cities exhibit positive indirect impacts, while only 41 cities show negative indirect impacts. The prevalent positive indirect effects can offset approximately 48% of the direct loss of tree coverage due to increased urban intensity, with full offsets achieved in Argentinian and arid regions of South America. In addition, human activity factors play the most important role in determining the indirect effects of urbanization on UTC coverage, followed by climatic and geographic factors. These findings will help us understand the impact of urbanization on UTC coverage along the urban intensity gradient and formulate policies and strategies to promote sustainable urban development in South America.

Understanding how buildings are distributed globally is crucial to revealing the human footprint on our home planet. This built environment affects local climate, land surface albedo, resource distribution, and many other key factors that influence well-being and human health. Despite this, quantitative and comprehensive data on the distribution and properties of buildings worldwide is lacking. To this end, by using a big data analytics approach and nearly 800,000 satellite images, we generated the highest resolution and highest accuracy building map ever created: the GlobalBuildingMap (GBM). A joint analysis of building maps and solar potentials indicates that rooftop solar energy can supply the global energy consumption need at a reasonable cost. Specifically, if solar panels were placed on the roofs of all buildings, they could supply 1.1-3.3 times – depending on the efficiency of the solar device – the global energy consumption in 2020, which is the year with the highest consumption on record. We also identified a clear geospatial correlation between building areas and key socioeconomic variables, which indicates our global building map can serve as an important input to modeling global socioeconomic needs and drivers.

In the remote sensing community, extracting buildings from remote sensing imagery has triggered great interest. While many studies have been conducted, a comprehensive review of these approaches that are applied to optical and synthetic aperture radar (SAR) imagery is still lacking. Therefore, we provide an in-depth review of both early efforts and recent advances, which are aimed at extracting geometrical structures or semantic attributes of buildings, including building footprint generation, building facade segmentation, roof segment and superstructure segmentation, building height retrieval, building-type classification, building change detection, and annotation data correction. Furthermore, a list of corresponding benchmark datasets is given. Finally, challenges and outlooks of existing approaches as well as promising applications are discussed to enhance comprehension within this realm of research.

Localizing desired objects from remote sensing images is of great use in practical applications. Referring image segmentation, which aims at segmenting out the objects to which a given expression refers, has been extensively studied in natural images. However, almost no research attention is given to this task of remote sensing imagery. Considering its potential for real-world applications, in this article, we introduce referring remote sensing image segmentation (RRSIS) to fill in this gap and make some insightful explorations. Specifically, we created a new dataset, called RefSegRS, for this task, enabling us to evaluate different methods. Afterward, we benchmark referring image segmentation methods of natural images on the RefSegRS dataset and find that these models show limited efficacy in detecting small and scattered objects. To alleviate this issue, we propose a language-guided cross-scale enhancement (LGCE) module that utilizes linguistic features to adaptively enhance multiscale visual features by integrating both deep and shallow features. The proposed dataset, benchmarking results, and the designed LGCE module provide insights into the design of a better RRSIS model.

The mass loss of glaciers outside the polar ice sheets has been accelerating during the past several decades and has been contributing to global sea-level rise. However, many of the mechanisms of this mass loss process are not well understood, especially the calving dynamics of marine-terminating glaciers, in part due to a lack of high-resolution calving front observations. Svalbard is an ideal site to study the climate sensitivity of glaciers as it is a region that has been undergoing amplified climate variability in both space and time compared to the global mean. Here we present a new high-resolution calving front dataset of 149 marine-terminating glaciers in Svalbard, comprising 124 919 glacier calving front positions during the period 1985–2023 (https://doi.org/10.5281/zenodo.10407266, Li et al., 2023). This dataset was generated using a novel automated deep-learning framework and multiple optical and SAR satellite images from Landsat, Terra-ASTER, Sentinel-2, and Sentinel-1 satellite missions. The overall calving front mapping uncertainty across Svalbard is 31 m. The newly derived calving front dataset agrees well with recent decadal calving front observations between 2000 and 2020 (Kochtitzky and Copland, 2022) and an annual calving front dataset between 2008 and 2022 (Moholdt et al., 2022). The calving fronts between our product and the latter deviate by 32±65m on average. The R2 of the glacier calving front change rates between these two products is 0.98, indicating an excellent match. Using this new calving front dataset, we identified widespread calving front retreats during the past four decades, across most regions in Svalbard except for a handful of glaciers draining the ice caps Vestfonna and Austfonna on Nordaustlandet. In addition, we identified complex patterns of glacier surging events overlaid with seasonal calving cycles. These data and findings provide insights into understanding glacier calving mechanisms and drivers. This new dataset can help improve estimates of glacier frontal ablation as a component of the integrated mass balance of marine-terminating glaciers.

In this article, we propose a multimodal co-learning framework for building change detection. This framework can be adopted to jointly train a Siamese bitemporal image network and a height difference (HDiff) network with labeled source data and unlabeled target data pairs. Three co-learning combinations (vanilla co-learning, fusion co-learning, and detached fusion co-learning) are proposed and investigated with two types of co-learning loss functions within our framework. Our experimental results demonstrate that the proposed methods are able to take advantage of unlabeled target data pairs and, therefore, enhance the performance of single-modal neural networks on the target data. In addition, our synthetic-to-real experiments demonstrate that the recently published synthetic dataset, Simulated Multimodal Aerial Remote Sensing (SMARS), is feasible to be used in real change detection scenarios, where the optimal result is with the F1 score of 79.29%.

In recent years, black-box machine learning approaches have become a dominant modeling paradigm for knowledge extraction in Remote Sensing. Despite the potential benefits of uncovering the inner workings of these models with explainable AI, a comprehensive overview summarizing the used explainable AI methods and their objectives, findings, and challenges in Remote Sensing applications is still missing. In this paper, we address this issue by performing a systematic review to identify the key trends of how explainable AI is used in Remote Sensing and shed light on novel explainable AI approaches and emerging directions that tackle specific Remote Sensing challenges. We also reveal the common patterns of explanation interpretation, discuss the extracted scientific insights in Remote Sensing, and reflect on the approaches used for explainable AI methods evaluation. Our review provides a complete summary of the state-of-the-art in the field. Further, we give a detailed outlook on the challenges and promising research directions, representing a basis for novel methodological development and a useful starting point for new researchers in the field of explainable AI in Remote Sensing.

Object detection (OD) is an essential and fundamental task in computer vision (CV) and satellite image processing. Existing deep learning methods have achieved impressive performance thanks to the availability of large-scale annotated datasets. Yet, in real-world applications, the availability of labels is limited. In this article, few-shot OD (FSOD) has emerged as a promising direction, which aims at enabling the model to detect novel objects with only few of them annotated. However, many existing FSOD algorithms overlook a critical issue: when an input image contains multiple novel objects and only a subset of them are annotated, the unlabeled objects will be considered as background during training. This can cause confusions and severely impact the model’s ability to recall novel objects. To address this issue, we propose a self-training-based FSOD (ST-FSOD) approach, which incorporates the self-training mechanism into the few-shot fine-tuning process. ST-FSOD aims to enable the discovery of novel objects that are not annotated and take them into account during training. On the one hand, we devise a two-branch region proposal networks (RPNs) to separate the proposal extraction of base and novel objects. On the another hand, we incorporate the student-teacher mechanism into RPN and the region-of-interest (RoI) head to include those highly confident yet unlabeled targets as pseudolabels. Experimental results demonstrate that our proposed method outperforms the state of the art in various FSOD settings by a large margin.

As extreme weather events become more frequent, understanding their impact on human health becomes increasingly crucial. However, the utilization of Earth Observation to effectively analyze the environmental context in relation to health remains limited. This limitation is primarily due to the lack of fine-grained spatial and temporal data in public and population health studies, hindering a comprehensive understanding of health outcomes. Additionally, obtaining appropriate environmental indices across different geographical levels and timeframes poses a challenge. For the years 2019 (pre-COVID) and 2020 (COVID), we collected spatio-temporal indicators for all Lower Layer Super Output Areas in England. These indicators included: i) 111 sociodemographic features linked to health in existing literature, ii) 43 environmental point features (e.g., greenery and air pollution levels), iii) 4 seasonal composite satellite images each with 11 bands, and iv) prescription prevalence associated with five medical conditions (depression, anxiety, diabetes, hypertension, and asthma), opioids and total prescriptions. We combined these indicators into a single MEDSAT dataset, the availability of which presents an opportunity for the machine learning community to develop new techniques specific to public health. These techniques would address challenges such as handling large and complex data volumes, performing effective feature engineering on environmental and sociodemographic factors, capturing spatial and temporal dependencies in the models, addressing imbalanced data distributions, developing novel computer vision methods for health modeling based on satellite imagery, ensuring model explainability, and achieving generalization beyond the specific geographical region.