HistoSeg++: Delving deeper with attention and multiscale feature fusion for biomarker segmentation

2026-07-02

AI Summary

This work addresses the limitations of existing biomarker segmentation methods in capturing multi-scale features and generalizing across datasets. To overcome these challenges, the authors propose an enhanced Nested-UNet architecture that innovatively integrates internal and external attention mechanisms with Squeeze-and-Excitation modules. This design dynamically emphasizes salient regions and recalibrates channel-wise features during upsampling. Additionally, an edge-aware loss function is introduced to improve boundary delineation accuracy. The proposed method substantially enhances multi-scale feature utilization and segmentation performance, consistently outperforming existing Nested-UNet variants across three public benchmark datasets and demonstrating superior cross-dataset generalization capability.

Abstract

Segmentation of biomarkers in medical images is frequently viewed as a first step towards medical image analysis in any bioinformatics or biomedical application. Despite progress, existing methods still struggle to capture information at multiple scales and to perform upsampling effectively across different datasets. These shortcomings often result in suboptimal generalization capabilities. Recently, architectures belonging to the Nested-UNet family excel in capturing multiscale contextual information and upsample them effectively. In this work, We propose a novel Nested-UNet architecture that effectively captures multi-scale contextual information. It includes inner and outer attention units to enhance focus during upsampling, along with channel-wise feature recalibration using squeeze-and-excitation modules, leading to improved segmentation performance. Additionally, the architecture integrates an edge-aware loss to emphasize boundary accuracy by assigning greater importance to edge regions. Tested extensively on three publicly available benchmark datasets. Our method demonstrates a generalization performance superior to existing Nested-UNet methods. Code: https://github.com/saadwazir/histosegplusplus

Problem

• biomarker segmentation
• multiscale feature fusion
• medical image analysis
• generalization capability
• upsampling

Innovation

• Nested-UNet
• attention mechanism
• multiscale feature fusion
• squeeze-and-excitation
• edge-aware loss