Tingying Peng

Sophia J. Wagner, Daniel Reisenbüchler, Nicholas P. West et al.

Background: Deep learning (DL) can extract predictive and prognostic biomarkers from routine pathology slides in colorectal cancer. For example, a DL test for the diagnosis of microsatellite instability (MSI) in CRC has been approved in 2022. Current approaches rely on convolutional neural networks (CNNs). Transformer networks are outperforming CNNs and are replacing them in many applications, but have not been used for biomarker prediction in cancer at a large scale. In addition, most DL approaches have been trained on small patient cohorts, which limits their clinical utility. Methods: In this study, we developed a new fully transformer-based pipeline for end-to-end biomarker prediction from pathology slides. We combine a pre-trained transformer encoder and a transformer network for patch aggregation, capable of yielding single and multi-target prediction at patient level. We train our pipeline on over 9,000 patients from 10 colorectal cancer cohorts. Results: A fully transformer-based approach massively improves the performance, generalizability, data efficiency, and interpretability as compared with current state-of-the-art algorithms. After training on a large multicenter cohort, we achieve a sensitivity of 0.97 with a negative predictive value of 0.99 for MSI prediction on surgical resection specimens. We demonstrate for the first time that resection specimen-only training reaches clinical-grade performance on endoscopic biopsy tissue, solving a long-standing diagnostic problem. Interpretation: A fully transformer-based end-to-end pipeline trained on thousands of pathology slides yields clinical-grade performance for biomarker prediction on surgical resections and biopsies. Our new methods are freely available under an open source license.

Daniel Reisenbüchler, Sophia J. Wagner, Melanie Boxberg et al.

Classical multiple instance learning (MIL) methods are often based on the identical and independent distributed assumption between instances, hence neglecting the potentially rich contextual information beyond individual entities. On the other hand, Transformers with global self-attention modules have been proposed to model the interdependencies among all instances. However, in this paper we question: Is global relation modeling using self-attention necessary, or can we appropriately restrict self-attention calculations to local regimes in large-scale whole slide images (WSIs)? We propose a general-purpose local attention graph-based Transformer for MIL (LA-MIL), introducing an inductive bias by explicitly contextualizing instances in adaptive local regimes of arbitrary size. Additionally, an efficiently adapted loss function enables our approach to learn expressive WSI embeddings for the joint analysis of multiple biomarkers. We demonstrate that LA-MIL achieves state-of-the-art results in mutation prediction for gastrointestinal cancer, outperforming existing models on important biomarkers such as microsatellite instability for colorectal cancer. Our findings suggest that local self-attention sufficiently models dependencies on par with global modules. Our LA-MIL implementation is available at https://github.com/agentdr1/LA_MIL.

Tomáš Chobola, Gesine Müller, Veit Dausmann et al.

The process of acquiring microscopic images in life sciences often results in image degradation and corruption, characterised by the presence of noise and blur, which poses significant challenges in accurately analysing and interpreting the obtained data. This paper proposes LUCYD, a novel method for the restoration of volumetric microscopy images that combines the Richardson-Lucy deconvolution formula and the fusion of deep features obtained by a fully convolutional network. By integrating the image formation process into a feature-driven restoration model, the proposed approach aims to enhance the quality of the restored images whilst reducing computational costs and maintaining a high degree of interpretability. Our results demonstrate that LUCYD outperforms the state-of-the-art methods in both synthetic and real microscopy images, achieving superior performance in terms of image quality and generalisability. We show that the model can handle various microscopy modalities and different imaging conditions by evaluating it on two different microscopy datasets, including volumetric widefield and light-sheet microscopy. Our experiments indicate that LUCYD can significantly improve resolution, contrast, and overall quality of microscopy images. Therefore, it can be a valuable tool for microscopy image restoration and can facilitate further research in various microscopy applications. We made the source code for the model accessible under https://github.com/ctom2/lucyd-deconvolution.

Florian Kofler, Johannes Wahle, Ivan Ezhov et al.

Machine learning models are typically evaluated by computing similarity with reference annotations and trained by maximizing similarity with such. Especially in the biomedical domain, annotations are subjective and suffer from low inter- and intra-rater reliability. Since annotations only reflect one interpretation of the real world, this can lead to sub-optimal predictions even though the model achieves high similarity scores. Here, the theoretical concept of PGT is introduced. PGT marks the point beyond which an increase in similarity with the \emph{reference annotation} stops translating to better RWMP. Additionally, a quantitative technique to approximate PGT by computing inter- and intra-rater reliability is proposed. Finally, four categories of PGT-aware strategies to evaluate and improve model performance are reviewed.

Tomáš Chobola, Yu Liu, Hanyi Zhang et al.

Current deep learning-based low-light image enhancement methods often struggle with high-resolution images, and fail to meet the practical demands of visual perception across diverse and unseen scenarios. In this paper, we introduce a novel approach termed CoLIE, which redefines the enhancement process through mapping the 2D coordinates of an underexposed image to its illumination component, conditioned on local context. We propose a reconstruction of enhanced-light images within the HSV space utilizing an implicit neural function combined with an embedded guided filter, thereby significantly reducing computational overhead. Moreover, we introduce a single image-based training loss function to enhance the model's adaptability to various scenes, further enhancing its practical applicability. Through rigorous evaluations, we analyze the properties of our proposed framework, demonstrating its superiority in both image quality and scene adaptability. Furthermore, our evaluation extends to applications in downstream tasks within low-light scenarios, underscoring the practical utility of CoLIE. The source code is available at https://github.com/ctom2/colie.

Yu Liu, Kurt Weiss, Nassir Navab et al.

Light-sheet fluorescence microscopy (LSFM) is a cutting-edge volumetric imaging technique that allows for three-dimensional imaging of mesoscopic samples with decoupled illumination and detection paths. Although the selective excitation scheme of such a microscope provides intrinsic optical sectioning that minimizes out-of-focus fluorescence background and sample photodamage, it is prone to light absorption and scattering effects, which results in uneven illumination and striping artifacts in the images adversely. To tackle this issue, in this paper, we propose a blind stripe artifact removal algorithm in LSFM, called DeStripe, which combines a self-supervised spatio-spectral graph neural network with unfolded Hessian prior. Specifically, inspired by the desirable properties of Fourier transform in condensing striping information into isolated values in the frequency domain, DeStripe firstly localizes the potentially corrupted Fourier coefficients by exploiting the structural difference between unidirectional stripe artifacts and more isotropic foreground images. Affected Fourier coefficients can then be fed into a graph neural network for recovery, with a Hessian regularization unrolled to further ensure structures in the standard image space are well preserved. Since in realistic, stripe-free LSFM barely exists with a standard image acquisition protocol, DeStripe is equipped with a Self2Self denoising loss term, enabling artifact elimination without access to stripe-free ground truth images. Competitive experimental results demonstrate the efficacy of DeStripe in recovering corrupted biomarkers in LSFM with both synthetic and real stripe artifacts.

Ziqi Yu, Xiaoyang Han, Shengjie Zhang et al.

Segmenting the fine structure of the mouse brain on magnetic resonance (MR) images is critical for delineating morphological regions, analyzing brain function, and understanding their relationships. Compared to a single MRI modality, multimodal MRI data provide complementary tissue features that can be exploited by deep learning models, resulting in better segmentation results. However, multimodal mouse brain MRI data is often lacking, making automatic segmentation of mouse brain fine structure a very challenging task. To address this issue, it is necessary to fuse multimodal MRI data to produce distinguished contrasts in different brain structures. Hence, we propose a novel disentangled and contrastive GAN-based framework, named MouseGAN++, to synthesize multiple MR modalities from single ones in a structure-preserving manner, thus improving the segmentation performance by imputing missing modalities and multi-modality fusion. Our results demonstrate that the translation performance of our method outperforms the state-of-the-art methods. Using the subsequently learned modality-invariant information as well as the modality-translated images, MouseGAN++ can segment fine brain structures with averaged dice coefficients of 90.0% (T2w) and 87.9% (T1w), respectively, achieving around +10% performance improvement compared to the state-of-the-art algorithms. Our results demonstrate that MouseGAN++, as a simultaneous image synthesis and segmentation method, can be used to fuse cross-modality information in an unpaired manner and yield more robust performance in the absence of multimodal data. We release our method as a mouse brain structural segmentation tool for free academic usage at https://github.com/yu02019.

Manuel Tran, Sophia J. Wagner, Melanie Boxberg et al.

In computational pathology, we often face a scarcity of annotations and a large amount of unlabeled data. One method for dealing with this is semi-supervised learning which is commonly split into a self-supervised pretext task and a subsequent model fine-tuning. Here, we compress this two-stage training into one by introducing S5CL, a unified framework for fully-supervised, self-supervised, and semi-supervised learning. With three contrastive losses defined for labeled, unlabeled, and pseudo-labeled images, S5CL can learn feature representations that reflect the hierarchy of distance relationships: similar images and augmentations are embedded the closest, followed by different looking images of the same class, while images from separate classes have the largest distance. Moreover, S5CL allows us to flexibly combine these losses to adapt to different scenarios. Evaluations of our framework on two public histopathological datasets show strong improvements in the case of sparse labels: for a H&E-stained colorectal cancer dataset, the accuracy increases by up to 9% compared to supervised cross-entropy loss; for a highly imbalanced dataset of single white blood cells from leukemia patient blood smears, the F1-score increases by up to 6%.

Yu Liu, Gesine Muller, Nassir Navab et al.

Light-sheet fluorescence microscopy (LSFM), a planar illumination technique that enables high-resolution imaging of samples, experiences defocused image quality caused by light scattering when photons propagate through thick tissues. To circumvent this issue, dualview imaging is helpful. It allows various sections of the specimen to be scanned ideally by viewing the sample from opposing orientations. Recent image fusion approaches can then be applied to determine in-focus pixels by comparing image qualities of two views locally and thus yield spatially inconsistent focus measures due to their limited field-of-view. Here, we propose BigFUSE, a global context-aware image fuser that stabilizes image fusion in LSFM by considering the global impact of photon propagation in the specimen while determining focus-defocus based on local image qualities. Inspired by the image formation prior in dual-view LSFM, image fusion is considered as estimating a focus-defocus boundary using Bayes Theorem, where (i) the effect of light scattering onto focus measures is included within Likelihood; and (ii) the spatial consistency regarding focus-defocus is imposed in Prior. The expectation-maximum algorithm is then adopted to estimate the focus-defocus boundary. Competitive experimental results show that BigFUSE is the first dual-view LSFM fuser that is able to exclude structured artifacts when fusing information, highlighting its abilities of automatic image fusion.

Tomáš Chobola, Gesine Müller, Veit Dausmann et al.

Non-blind deconvolution aims to restore a sharp image from its blurred counterpart given an obtained kernel. Existing deep neural architectures are often built based on large datasets of sharp ground truth images and trained with supervision. Sharp, high quality ground truth images, however, are not always available, especially for biomedical applications. This severely hampers the applicability of current approaches in practice. In this paper, we propose a novel non-blind deconvolution method that leverages the power of deep learning and classic iterative deconvolution algorithms. Our approach combines a pre-trained network to extract deep features from the input image with iterative Richardson-Lucy deconvolution steps. Subsequently, a zero-shot optimisation process is employed to integrate the deconvolved features, resulting in a high-quality reconstructed image. By performing the preliminary reconstruction with the classic iterative deconvolution method, we can effectively utilise a smaller network to produce the final image, thus accelerating the reconstruction whilst reducing the demand for valuable computational resources. Our method demonstrates significant improvements in various real-world applications non-blind deconvolution tasks.

Tomas Chobola, Anton Theileis, Jan Taucher et al.

We present a model for non-blind image deconvolution that incorporates the classic iterative method into a deep learning application. Instead of using large over-parameterised generative networks to create sharp picture representations, we build our network based on the iterative Landweber deconvolution algorithm, which is integrated with trainable convolutional layers to enhance the recovered image structures and details. Additional to the data fidelity term, we also add Hessian and sparse constraints as regularization terms to improve the image reconstruction quality. Our proposed model is \textit{self-supervised} and converges to a solution based purely on the input blurred image and respective blur kernel without the requirement of any pre-training. We evaluate our technique using standard computer vision benchmarking datasets as well as real microscope images obtained by our enhanced depth-of-field (EDOF) underwater microscope, demonstrating the capabilities of our model in a real-world application. The quantitative results demonstrate that our approach is competitive with state-of-the-art non-blind image deblurring methods despite having a fraction of the parameters and not being pre-trained, demonstrating the efficiency and efficacy of embedding a classic deconvolution approach inside a deep network.

Valentin Koch, Sophia J. Wagner, Salome Kazeminia et al.

In hematology, computational models offer significant potential to improve diagnostic accuracy, streamline workflows, and reduce the tedious work of analyzing single cells in peripheral blood or bone marrow smears. However, clinical adoption of computational models has been hampered by the lack of generalization due to large batch effects, small dataset sizes, and poor performance in transfer learning from natural images. To address these challenges, we introduce DinoBloom, the first foundation model for single cell images in hematology, utilizing a tailored DINOv2 pipeline. Our model is built upon an extensive collection of 13 diverse, publicly available datasets of peripheral blood and bone marrow smears, the most substantial open-source cohort in hematology so far, comprising over 380,000 white blood cell images. To assess its generalization capability, we evaluate it on an external dataset with a challenging domain shift. We show that our model outperforms existing medical and non-medical vision models in (i) linear probing and k-nearest neighbor evaluations for cell-type classification on blood and bone marrow smears and (ii) weakly supervised multiple instance learning for acute myeloid leukemia subtyping by a large margin. A family of four DinoBloom models (small, base, large, and giant) can be adapted for a wide range of downstream applications, be a strong baseline for classification problems, and facilitate the assessment of batch effects in new datasets. All models are available at github.com/marrlab/DinoBloom.

Mikhail Konov, Lion J. Gleiter, Khoa Co et al.

AI tools can greatly enhance the analysis of organoid microscopy images, from detection and segmentation to feature extraction and classification. However, their limited accessibility to biologists without programming experience remains a major barrier, resulting in labor-intensive and largely manual workflows. Although a few AI models for organoid analysis have been developed, most existing tools remain narrowly focused on specific tasks. In this work, we introduce the Napari Organoid Analyzer (NOA), a general purpose graphical user interface to simplify AI-based organoid analysis. NOA integrates modules for detection, segmentation, tracking, feature extraction, custom feature annotation and ML-based feature prediction. It interfaces multiple state-of-the-art algorithms and is implemented as an open-source napari plugin for maximal flexibility and extensibility. We demonstrate the versatility of NOA through three case studies, involving the quantification of morphological changes during organoid differentiation, assessment of phototoxicity effects, and prediction of organoid viability and differentiation state. Together, these examples illustrate how NOA enables comprehensive, AI-driven organoid image analysis within an accessible and extensible framework.

Ziqi Yu, Botao Zhao, Shengjie Zhang et al.

Synthesizing medical images while preserving their structural information is crucial in medical research. In such scenarios, the preservation of anatomical content becomes especially important. Although recent advances have been made by incorporating instance-level information to guide translation, these methods overlook the spatial coherence of structural-level representation and the anatomical invariance of content during translation. To address these issues, we introduce hierarchical granularity discrimination, which exploits various levels of semantic information present in medical images. Our strategy utilizes three levels of discrimination granularity: pixel-level discrimination using a Brain Memory Bank, structure-level discrimination on each brain structure with a re-weighting strategy to focus on hard samples, and global-level discrimination to ensure anatomical consistency during translation. The image translation performance of our strategy has been evaluated on three independent datasets (UK Biobank, IXI, and BraTS 2018), and it has outperformed state-of-the-art algorithms. Particularly, our model excels not only in synthesizing normal structures but also in handling abnormal (pathological) structures, such as brain tumors, despite the variations in contrast observed across different imaging modalities due to their pathological characteristics. The diagnostic value of synthesized MR images containing brain tumors has been evaluated by radiologists. This indicates that our model may offer an alternative solution in scenarios where specific MR modalities of patients are unavailable. Extensive experiments further demonstrate the versatility of our method, providing unique insights into medical image translation.

Oytun Demirbilek, Tingying Peng, Alaa Bessadok

A morphological brain graph depicting a connectional fingerprint is of paramount importance for charting brain dysconnectivity patterns. Such data often has missing observations due to various reasons such as time-consuming and incomplete neuroimage processing pipelines. Thus, predicting a target brain graph from a source graph is crucial for better diagnosing neurological disorders with minimal data acquisition resources. Many brain graph generative models were proposed for promising results, yet they are mostly based on generative adversarial networks (GAN), which could suffer from mode collapse and require large training datasets. Recent developments in diffusion models address these problems by offering essential properties such as a stable training objective and easy scalability. However, applying a diffusion process to graph edges fails to maintain the topological symmetry of the brain connectivity matrices. To meet these challenges, we propose the Graph Residual Noise Learner Network (Grenol-Net), the first graph diffusion model for predicting a target graph from a source graph.

Joel Valdivia Ortega, Tingying Peng, Marion Jasnin

Semantic segmentation is essential for analysing anatomical features in biomedical research, yet a performance gap remains for Vision Transformers (ViTs) in the field, particularly for sparse, fine-structured, and low signal-to-noise targets. We attribute this challenge in part to the lightweight pixel decoders commonly used in promptable ViT models, who may lack the local inductive bias needed for high-precision biomedical masks. We bridge this gap by introducing ViTC-UNet, which conditions a UNet on frozen pre-trained ViT representations through learnable tokens and a two-way attention decoder. This combines ViT global visual priors with the local inductive bias and high-resolution decoding capacity of UNets, while avoiding end-to-end ViT fine-tuning even in cross-domain settings. ViTC-UNet outperforms baseline results in semantic segmentation tasks across MRI and CT modalities, demonstrating that structure-conditioned UNet decoding can efficiently adapt large-scale visual priors to high-complexity biomedical segmentation.

Tong Ding, Sophia J. Wagner, Andrew H. Song et al.

The field of computational pathology has been transformed with recent advances in foundation models that encode histopathology region-of-interests (ROIs) into versatile and transferable feature representations via self-supervised learning (SSL). However, translating these advancements to address complex clinical challenges at the patient and slide level remains constrained by limited clinical data in disease-specific cohorts, especially for rare clinical conditions. We propose TITAN, a multimodal whole slide foundation model pretrained using 335,645 WSIs via visual self-supervised learning and vision-language alignment with corresponding pathology reports and 423,122 synthetic captions generated from a multimodal generative AI copilot for pathology. Without any finetuning or requiring clinical labels, TITAN can extract general-purpose slide representations and generate pathology reports that generalize to resource-limited clinical scenarios such as rare disease retrieval and cancer prognosis. We evaluate TITAN on diverse clinical tasks and find that TITAN outperforms both ROI and slide foundation models across machine learning settings such as linear probing, few-shot and zero-shot classification, rare cancer retrieval and cross-modal retrieval, and pathology report generation.

Benedikt Roth, Valentin Koch, Sophia J. Wagner et al.

To handle the large scale of whole slide images in computational pathology, most approaches first tessellate the images into smaller patches, extract features from these patches, and finally aggregate the feature vectors with weakly-supervised learning. The performance of this workflow strongly depends on the quality of the extracted features. Recently, foundation models in computer vision showed that leveraging huge amounts of data through supervised or self-supervised learning improves feature quality and generalizability for a variety of tasks. In this study, we benchmark the most popular vision foundation models as feature extractors for histopathology data. We evaluate the models in two settings: slide-level classification and patch-level classification. We show that foundation models are a strong baseline. Our experiments demonstrate that by finetuning a foundation model on a single GPU for only two hours or three days depending on the dataset, we can match or outperform state-of-the-art feature extractors for computational pathology. These findings imply that even with little resources one can finetune a feature extractor tailored towards a specific downstream task and dataset. This is a considerable shift from the current state, where only few institutions with large amounts of resources and datasets are able to train a feature extractor. We publish all code used for training and evaluation as well as the finetuned models.

Rushin H. Gindra, Giovanni Palla, Mathias Nguyen et al.

Spatial transcriptomics enables simultaneous measurement of gene expression and tissue morphology, offering unprecedented insights into cellular organization and disease mechanisms. However, the field lacks comprehensive benchmarks for evaluating multimodal learning methods that leverage both histology images and gene expression data. Here, we present HESCAPE, a large-scale benchmark for cross-modal contrastive pretraining in spatial transcriptomics, built on a curated pan-organ dataset spanning 6 different gene panels and 54 donors. We systematically evaluated state-of-the-art image and gene expression encoders across multiple pretraining strategies and assessed their effectiveness on two downstream tasks: gene mutation classification and gene expression prediction. Our benchmark demonstrates that gene expression encoders are the primary determinant of strong representational alignment, and that gene models pretrained on spatial transcriptomics data outperform both those trained without spatial data and simple baseline approaches. However, downstream task evaluation reveals a striking contradiction: while contrastive pretraining consistently improves gene mutation classification performance, it degrades direct gene expression prediction compared to baseline encoders trained without cross-modal objectives. We identify batch effects as a key factor that interferes with effective cross-modal alignment. Our findings highlight the critical need for batch-robust multimodal learning approaches in spatial transcriptomics. To accelerate progress in this direction, we release HESCAPE, providing standardized datasets, evaluation protocols, and benchmarking tools for the community

Joel Valdivia Ortega, Lorenz Lamm, Franziska Eckardt et al.

Vision Transformers (ViTs), such as DINOv2, achieve strong performance across domains but often repurpose low-informative patch tokens in ways that reduce the interpretability of attention and feature maps. This challenge is especially evident in medical imaging, where domain shifts can degrade both performance and transparency. In this paper, we introduce Randomized-MLP (RMLP) regularization, a contrastive learning-based method that encourages more semantically aligned representations. We use RMLPs when fine-tuning DINOv2 to both medical and natural image modalities, showing that it improves or maintains downstream performance while producing more interpretable attention maps. We also provide a mathematical analysis of RMLPs, offering insights into its role in enhancing ViT-based models and advancing our understanding of contrastive learning.

Tomáš Chobola, Julia A. Schnabel, Tingying Peng

Current self-supervised denoising techniques achieve impressive results, yet their real-world application is frequently constrained by substantial computational and memory demands, necessitating a compromise between inference speed and reconstruction quality. In this paper, we present an ultra-lightweight model that addresses this challenge, achieving both fast denoising and high quality image restoration. Built upon the Noise2Noise training framework-which removes the reliance on clean reference images or explicit noise modeling-we introduce an innovative multistage denoising pipeline named Noise2Detail (N2D). During inference, this approach disrupts the spatial correlations of noise patterns to produce intermediate smooth structures, which are subsequently refined to recapture fine details directly from the noisy input. Extensive testing reveals that Noise2Detail surpasses existing dataset-free techniques in performance, while requiring only a fraction of the computational resources. This combination of efficiency, low computational cost, and data-free approach make it a valuable tool for biomedical imaging, overcoming the challenges of scarce clean training data-due to rare and complex imaging modalities-while enabling fast inference for practical use.

Manuel Tran, Amal Lahiani, Yashin Dicente Cid et al.

Vision Transformers (ViTs) and Swin Transformers (Swin) are currently state-of-the-art in computational pathology. However, domain experts are still reluctant to use these models due to their lack of interpretability. This is not surprising, as critical decisions need to be transparent and understandable. The most common approach to understanding transformers is to visualize their attention. However, attention maps of ViTs are often fragmented, leading to unsatisfactory explanations. Here, we introduce a novel architecture called the B-cos Vision Transformer (BvT) that is designed to be more interpretable. It replaces all linear transformations with the B-cos transform to promote weight-input alignment. In a blinded study, medical experts clearly ranked BvTs above ViTs, suggesting that our network is better at capturing biomedically relevant structures. This is also true for the B-cos Swin Transformer (Bwin). Compared to the Swin Transformer, it even improves the F1-score by up to 4.7% on two public datasets.

Manuel Tran, Yashin Dicente Cid, Amal Lahiani et al.

Training multimodal foundation models is challenging due to the limited availability of multimodal datasets. While many public datasets pair images with text, few combine images with audio or text with audio. Even rarer are datasets that align all three modalities at once. Critical domains such as healthcare, infrastructure, or transportation are particularly affected by missing modalities. This makes it difficult to integrate all modalities into a large pre-trained neural network that can be used out-of-the-box or fine-tuned for different downstream tasks. We introduce LoReTTa (Linking mOdalities with a tRansitive and commutativE pre-Training sTrAtegy) to address this understudied problem. Our self-supervised framework unifies causal modeling and masked modeling with the rules of commutativity and transitivity. This allows us to transition within and between modalities. As a result, our pre-trained models are better at exploring the true underlying joint probability distribution. Given a dataset containing only the disjoint combinations (A, B) and (B, C), LoReTTa can model the relation A <-> C with A <-> B <-> C. In particular, we show that a transformer pre-trained with LoReTTa can handle any mixture of modalities at inference time, including the never-seen pair (A, C) and the triplet (A, B, C). We extensively evaluate our approach on a synthetic, medical, and reinforcement learning dataset. Across different domains, our universal multimodal transformer consistently outperforms strong baselines such as GPT, BERT, and CLIP on tasks involving the missing modality tuple.

Ye Liu, Sophia J. Wagner, Tingying Peng

Annotating microscopy images for nuclei segmentation is laborious and time-consuming. To leverage the few existing annotations, also across multiple modalities, we propose a novel microscopy-style augmentation technique based on a generative adversarial network (GAN). Unlike other style transfer methods, it can not only deal with different cell assay types and lighting conditions, but also with different imaging modalities, such as bright-field and fluorescence microscopy. Using disentangled representations for content and style, we can preserve the structure of the original image while altering its style during augmentation. We evaluate our data augmentation on the 2018 Data Science Bowl dataset consisting of various cell assays, lighting conditions, and imaging modalities. With our style augmentation, the segmentation accuracy of the two top-ranked Mask R-CNN-based nuclei segmentation algorithms in the competition increases significantly. Thus, our augmentation technique renders the downstream task more robust to the test data heterogeneity and helps counteract class imbalance without resampling of minority classes.

Sophia J. Wagner, Nadieh Khalili, Raghav Sharma et al.

In digital pathology, different staining procedures and scanners cause substantial color variations in whole-slide images (WSIs), especially across different laboratories. These color shifts result in a poor generalization of deep learning-based methods from the training domain to external pathology data. To increase test performance, stain normalization techniques are used to reduce the variance between training and test domain. Alternatively, color augmentation can be applied during training leading to a more robust model without the extra step of color normalization at test time. We propose a novel color augmentation technique, HistAuGAN, that can simulate a wide variety of realistic histology stain colors, thus making neural networks stain-invariant when applied during training. Based on a generative adversarial network (GAN) for image-to-image translation, our model disentangles the content of the image, i.e., the morphological tissue structure, from the stain color attributes. It can be trained on multiple domains and, therefore, learns to cover different stain colors as well as other domain-specific variations introduced in the slide preparation and imaging process. We demonstrate that HistAuGAN outperforms conventional color augmentation techniques on a classification task on the publicly available dataset Camelyon17 and show that it is able to mitigate present batch effects.

Ario Sadafi, Asya Makhro, Anna Bogdanova et al.

Red blood cells are highly deformable and present in various shapes. In blood cell disorders, only a subset of all cells is morphologically altered and relevant for the diagnosis. However, manually labeling of all cells is laborious, complicated and introduces inter-expert variability. We propose an attention based multiple instance learning method to classify blood samples of patients suffering from blood cell disorders. Cells are detected using an R-CNN architecture. With the features extracted for each cell, a multiple instance learning method classifies patient samples into one out of four blood cell disorders. The attention mechanism provides a measure of the contribution of each cell to the overall classification and significantly improves the network's classification accuracy as well as its interpretability for the medical expert.

Gerda Bortsova, Gijs van Tulder, Florian Dubost et al.

Intracranial carotid artery calcification (ICAC) is a major risk factor for stroke, and might contribute to dementia and cognitive decline. Reliance on time-consuming manual annotation of ICAC hampers much demanded further research into the relationship between ICAC and neurological diseases. Automation of ICAC segmentation is therefore highly desirable, but difficult due to the proximity of the lesions to bony structures with a similar attenuation coefficient. In this paper, we propose a method for automatic segmentation of ICAC; the first to our knowledge. Our method is based on a 3D fully convolutional neural network that we extend with two regularization techniques. Firstly, we use deep supervision (hidden layers supervision) to encourage discriminative features in the hidden layers. Secondly, we augment the network with skip connections, as in the recently developed ResNet, and dropout layers, inserted in a way that skip connections circumvent them. We investigate the effect of skip connections and dropout. In addition, we propose a simple problem-specific modification of the network objective function that restricts the focus to the most important image regions and simplifies the optimization. We train and validate our model using 882 CT scans and test on 1,000. Our regularization techniques and objective improve the average Dice score by 7.1%, yielding an average Dice of 76.2% and 97.7% correlation between predicted ICAC volumes and manual annotations.

Gerda Bortsova, Michael Sterr, Lichao Wang et al.

Intestinal enteroendocrine cells secrete hormones that are vital for the regulation of glucose metabolism but their differentiation from intestinal stem cells is not fully understood. Asymmetric stem cell divisions have been linked to intestinal stem cell homeostasis and secretory fate commitment. We monitored cell divisions using 4D live cell imaging of cultured intestinal crypts to characterize division modes by means of measurable features such as orientation or shape. A statistical analysis of these measurements requires annotation of mitosis events, which is currently a tedious and time-consuming task that has to be performed manually. To assist data processing, we developed a learning based method to automatically detect mitosis events. The method contains a dual-phase framework for joint detection of dividing cells (mothers) and their progeny (daughters). In the first phase we detect mother and daughters independently using Hough Forest whilst in the second phase we associate mother and daughters by modelling their joint probability as Conditional Random Field (CRF). The method has been evaluated on 32 movies and has achieved an AUC of 72%, which can be used in conjunction with manual correction and dramatically speed up the processing pipeline.