Superpixel Boundary Correction for Weakly-Supervised Semantic Segmentation on Histopathology Images

2025-01-07

AI Summary

In weakly supervised semantic segmentation, Class Activation Mapping (CAM) suffers from blurred boundaries and low spatial resolution, leading to inaccurate delineation of the tumor microenvironment. To address this, we propose a multi-level superpixel boundary refinement algorithm: (1) SLIC superpixel structural priors are first integrated into CAM optimization; (2) flood-fill-based boundary propagation is employed to refine coarse CAMs; and (3) multi-scale feature fusion enhances localization accuracy. Our method significantly improves boundary clarity and consistency between breast cancer epithelial tissue and stroma in histopathological images. Evaluated on a breast cancer pathology dataset, it achieves 71.08% mean Intersection-over-Union (mIoU), substantially outperforming baseline CAM approaches. The framework delivers both high segmentation accuracy and clinically interpretable outputs—enabling precise, boundary-aware tumor microenvironment characterization without requiring pixel-level annotations.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Algorithm Configuration

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

With the rapid advancement of deep learning, computational pathology has made significant progress in cancer diagnosis and subtyping. Tissue segmentation is a core challenge, essential for prognosis and treatment decisions. Weakly supervised semantic segmentation (WSSS) reduces the annotation requirement by using image-level labels instead of pixel-level ones. However, Class Activation Map (CAM)-based methods still suffer from low spatial resolution and unclear boundaries. To address these issues, we propose a multi-level superpixel correction algorithm that refines CAM boundaries using superpixel clustering and floodfill. Experimental results show that our method achieves great performance on breast cancer segmentation dataset with mIoU of 71.08%, significantly improving tumor microenvironment boundary delineation.

Problem

• Weakly Supervised Semantic Segmentation
• Cellular Image Analysis
• Cancer Diagnosis

Innovation

• Weakly Supervised Semantic Segmentation
• Boundary Optimization
• Breast Cancer Tissue Identification