Problem
Research questions and friction points this paper is trying to address.
weakly supervised segmentation
histopathology
boundary ambiguity
pseudo-mask generation
image-level annotation
ProBAG: Prototype-Guided Boundary-Aware Graph Diffusion for Weakly Supervised Histopathology Segmentation
🤖 AI Summary
This work addresses the limitations of conventional Class Activation Mapping (CAM) approaches in weakly supervised semantic segmentation of histopathology images using only image-level labels—specifically, their tendency to produce blurry boundaries and over-localize on highly discriminative regions. To overcome these issues, the authors propose ProBAG, a novel method that integrates pathology-aligned CONCH text prototypes with multi-scale frozen UNI visual features to generate high-quality pseudo-masks. ProBAG further introduces class-wise power recalibration to preserve total foreground activation and designs an attention-based contextual discrepancy-driven graph diffusion mechanism to enhance boundary awareness without relying on CRF or external models. Integrated within a Phikon-FPN two-stage segmentation framework, ProBAG achieves state-of-the-art performance on BCSS-WSSS and LUAD-HistoSeg benchmarks. Ablation studies confirm that pathological text semantics provide the largest contribution, while graph diffusion offers complementary gains.
📝 Abstract
Weakly supervised semantic segmentation enables histopathology tissue segmentation from image-level annotations, avoiding costly pixel-level labeling by expert pathologists. However, CAM-based methods often localize only highly discriminative regions and remain unreliable near tissue interfaces. We propose ProBAG, a stage-1 pseudo-mask generator that combines dataset-specific visual prototypes with pathology-aligned CONCH text prototypes over multi-scale frozen UNI features. ProBAG introduces two complementary mechanisms: class-wise power recalibration that reshapes inter-class competition while preserving the total foreground activation mass at each pixel, and one-step graph diffusion in which feature affinities are penalized by a late-transformer attention-context discrepancy used as a soft structural boundary cue. The resulting stage-1 pseudo-masks require neither CRF nor an external segmentation model; for complete two-stage comparison, they additionally supervise a downstream Phikon-FPN segmenter. Experiments on BCSS-WSSS and LUAD-HistoSeg show consistent gains over recent WSSS approaches, while ablations indicate that pathology-aligned text semantics provide the largest improvement and graph refinement provides a smaller complementary gain. The code is available at: https://github.com/wterrr/WSSS
Innovation
Methods, ideas, or system contributions that make the work stand out.
Prototype-Guided
Boundary-Aware Graph Diffusion
Weakly Supervised Segmentation
Pathology-Aligned Text Semantics
Power Recalibration
🔎 Similar Papers
No similar papers found.
