Utilizing Weak-to-Strong Consistency for Semi-Supervised Glomeruli Segmentation

TL;DR

This paper introduces a semi-supervised learning method leveraging weak-to-strong consistency for more accurate glomeruli segmentation in renal biopsy images, addressing annotation challenges and variability across datasets.

Contribution

The proposed approach is validated on multiple real-world datasets, demonstrating superior performance over existing supervised models like U-Net and SegFormer.

Findings

Outperforms baseline models in segmentation accuracy

Effective on multiple real-world datasets

Reduces reliance on extensive annotations

Abstract

Accurate segmentation of glomerulus instances attains high clinical significance in the automated analysis of renal biopsies to aid in diagnosing and monitoring kidney disease. Analyzing real-world histopathology images often encompasses inter-observer variability and requires a labor-intensive process of data annotation. Therefore, conventional supervised learning approaches generally achieve sub-optimal performance when applied to external datasets. Considering these challenges, we present a semi-supervised learning approach for glomeruli segmentation based on the weak-to-strong consistency framework validated on multiple real-world datasets. Our experimental results on 3 independent datasets indicate superior performance of our approach as compared with existing supervised baseline models such as U-Net and SegFormer.