Through the Schrödinger Bridge: Benchmarking Antemortem Image Restoration from Postmortem Autolysis to Enhance Forensic Diagnostics
Authors: Shuang Hao, Jiacheng Yue, Yaxuan Zhao, Fan Wang, Jianhua Ma, Erwen Huang, Chunfeng Lian
Organizations: Key Laboratory of Biomedical Information Engineering of Ministry of Education, School of Life Science and Technology, Xi’an Jiaotong University, Xi’an, China · Research Center for Intelligent Medical Equipment and Devices (IMED), Xi’an Jiaotong University, Xi’an 710049, China · Faculty of Forensic Medicine, Zhongshan School of Medicine, Sun Yat-Sen University, Guangzhou, 510080, China · School of Mathematics and Statistics, Xi’an Jiaotong University, Xi’an, China
Forensic histopathology, essential for determining cause of death and disease diagnosis, is severely impeded by postmortem autolysis, i.e., an irreversible, stochastic degradation process that distorts tissue morphology and introduces diagnostic subjectivity, thereby underscoring the value of restoring autolyzed images to a diagnostically plausible, pre-autolysis state for improving objectivity in forensic practice. This restoration task is fundamentally challenging due to the large, non-deterministic morphological changes caused by autolysis and the infeasibility of pixel-wise paired data, which invalidates assumptions underlying supervised and cycle/structure-consistent unpaired translation methods. To address this, we formalize forensic histopathology autolysis restoration as a new task: under unpaired supervision, transform postmortem images with severe autolysis into diagnostically meaningful ``antemortem'' representations. We contribute AutoPath, the first homologous yet unpaired dataset for this problem, constructed by splitting specimens into adjacent tissue blocks---one processed immediately, the other exposed to induce autolysis---yielding nearly ten thousand 10× patches from 69 cases with varying liver conditions. We further frame the problem as a Schrödinger Bridge between the autolyzed and non-autolyzed distributions, offering a principled approach to modeling stochastic, severe morphological degradation. Critically, we demonstrate the misalignment of generic image-level generative metrics (e.g., FID) with diagnostic utility and propose a forensically grounded, slide-level diagnostic distribution consistency evaluation. Overall, this work establishes a reproducible benchmark (encompassing task definition, a real-world dataset, and an evaluation methodology) toward rigorous and practically meaningful progress in autolysis restoration for forensic pathology.
Figures & tables
Figure 1: Motivation and dataset overview. (a) As time elapses after exposure, tissue undergoes autolysis that disrupts morphological cues, making expert diagnosis less reliable compared with fresh slides. (b) Composition of our liver pathology dataset, showing the distribution of cases by sex/age groups and their co-occurring pathology tags (chord diagram), together with overall counts of subjects and image patches.
Figure 2: Overview of the proposed framework. Left: Dataset pipeline. Fresh liver tissues are divided into adjacent blocks, where one block is immediately processed into a non-autolyzed slide and the other is exposed to ambient conditions for seven days to induce autolysis, followed by patch extraction and filtering. Top-right: Schrödinger Bridge-based restoration model. Autolyzed patches sampled from π0 are progressively transported to the non-autolyzed domain π1 via multi-step stochastic refinement. Bottom-right: Slide-level evaluation framework. Generated patches and real non-autolyzed patches are classified to obtain diagnostic category distributions, and slide-level proportion consistency is used to assess diagnostic validity.
Method
Patch-level
Slide-level (69 cases)
FID ↓
Expert Val. ↑
WSDC ↑
WKL ↓
High-Conf ↑
CycleGAN
51.24
18.00%
0.748
0.406
0.885
DN
51.87
2.38%
0.723
0.378
0.840
KIN
39.78
18.75%
0.740
0.367
0.830
CUT
53.46
17.65%
0.769
0.399
0.889
Ours
55.63
23.81%
0.799
0.343
0.897
Table 1: Quantitative and slide-level evaluation on AutoPath. We report patch-level FID (lower is better) with expert-based validation (higher is better), and slide-level diagnostic distribution consistency on 69 cases. Higher is better for WSDC and High-Conf Ratio; lower is better for WKL.
Figure 3: Qualitative comparison of autolysis restoration on AutoPath. We visualize representative autolyzed liver patches and the corresponding restoration results produced by different methods. Existing approaches often retain or amplify autolysis-induced artifacts (e.g., blurred nuclei and disrupted tissue organization), whereas Schrödinger Bridge-based model yields more plausible cellular morphology and tissue architecture according to forensic pathology experts.
Synthetic histopathology image generation addresses patient-privacy concerns and the growing data demands of foundation models. Existing state-of-the-art histopathology generative models use pretrained Vision Foundation Models (VFMs) as conditioning signals. We show this yields conditioning-dominated diversity: on TCGA-BRCA, 62-75% of their output diversity is attributable to the conditioning signal rather than the learned latent space, while de novo synthesis still requires a VFM at inference. We instead use histopathology VFMs as the latent space itself: their patch tokens are ℓ2-normalized on the unit hypersphere Sd−1 with strong angular dominance and intrinsic curvature, motivating a Riemannian formulation. We present STREAM, the first framework to apply Riemannian flow matching in the histopathology domain, in two stages: 1) a bridge-type stochastic perturbation that establishes per-token rectifiability on Sd−1 for training a Diffusion Transformer, and 2) a novel decoder training design whose noise covariance is anisotropic in the left-singular basis of the per-token tangent-projected velocity-field Jacobian, spending a large robustness budget on its low-response directions and a small one on its high-response directions. Across TCGA-BRCA and TCGA-COADREAD, STREAM achieves state-of-the-art gFID and ranks first on nearly all histopathology-specific metrics as well. Code and a public gallery of generated images are available at https://chokevin8.github.io/STREAM-Patho/.
Although deep learning has advanced Whole Slide Image (WSI) Analysis, tissue artifacts like bubbles and folds often cause silent failures by concealing essential morphology. Current pathology image restoration methods are mostly restricted to small patches, struggling to maintain global structural coherence at a megapixel scale. We introduce RestorePath, a framework for globally consistent megapixel scale inpainting that reconstructs diagnostic structures in histological image to prevent incorrect high-confidence predictions and lower error rates. Our model utilizes a Latent Diffusion Model (LDM) conditioned on Pathology Foundation Model (PFM) embeddings, integrating Large Kernel Attention (LKA) to manage long-range dependencies during random patch diffusion. Enhanced by Distance-Weighted Interpolation (DWI) and an Adaptive Guidance Scale (AGS), RestorePath ensures structural consistency and fidelity by modulating information from surrounding patches. Evaluations across TCGA-BRCA, BACH, and Camelyon16 datasets for images ranging from 512 to 4608 pixels demonstrate state-of-the-art performance in maintaining histological consistency. RestorePath significantly improves downstream Computational Pathology (CP) tasks, outperforming both raw artifact images and the conventional Detect-and-Discard (D&D) approach. The code is available at https://github.com/PathfinderLab/RestorePath
Hyeseong Lee, Eunsu Kim, D M Bappy +6
Department of Biomedical Informatics, Korea University College of Medicine, Seoul, Korea · Department of Pathology, Korea University Anam Hospital, College of Medicine, Korea University, Seoul, Korea · Seegene Medical Foundation, Seoul, Korea +3
Synthetic histopathology image generation has emerged as an approach that may address data scarcity in computational pathology, yet current evaluation methodologies may not fully assess synthetic data quality for medical applications. This work investigates and addresses limitations in existing evaluation metrics, investigating an approach for assessing synthetic histopathology image quality through domain-specific metrics and downstream task validation. We show that conventional synthetic data evaluation metrics such as Frechet Inception Distance (FID) and Inception Score (IS) may have limitations when applied to histopathology images due to their reliance on ImageNet-pretrained feature extractors. To address these limitations, we propose for consideration modified FID and IS approaches utilizing foundation models pretrained on digital pathology datasets, supplemented by precision-recall based metrics as part of an additional quality assessment. Using conditional denoising diffusion models trained on four benchmark datasets, with a two-step training approach, we generated synthetic datasets with systematically varied quality characteristics. We also measured the correlation between the synthetic data quality metrics with downstream nuclei segmentation performance using common metrics including the aggregated Jaccard index (AJI+) and the Dice coefficient. The study results suggest that pathology-specific metrics may provide improved discriminative power. Specifically, the modified Inception Score indicates higher correlation with downstream task performance (r=0.6096 with AJI+, p=0.0122), compared to the original IS (r=0.0708, p=0.7944). Our observations indicate that increasing the variety of generated training data has a higher positive correlation with segmentation model performance than improving the visual fidelity of individual generated images.
Seyed Kahaki, Shijie Li, Weijie Chen +1
Division of Imaging, Diagnostics, and Software Reliability, U.S. Food and Drug Administration (FDA)