Enhancing Weakly Supervised Semantic Segmentation for Fibrosis via Controllable Image Generation

TL;DR

This paper introduces DiffSeg, a weakly supervised segmentation method that uses diffusion-based image generation to improve fibrosis detection in HRCT scans, reducing manual labeling effort.

Contribution

The paper presents a novel diffusion-based generative approach integrated with WSSS to enhance fibrosis segmentation accuracy from image-level annotations.

Findings

Significant improvement in pseudo mask accuracy.

Reduced manual annotation complexity.

Enhanced consistency of segmentation masks.

Abstract

Fibrotic Lung Disease (FLD) is a severe condition marked by lung stiffening and scarring, leading to respiratory decline. High-resolution computed tomography (HRCT) is critical for diagnosing and monitoring FLD; however, fibrosis appears as irregular, diffuse patterns with unclear boundaries, leading to high inter-observer variability and time-intensive manual annotation. To tackle this challenge, we propose DiffSeg, a novel weakly supervised semantic segmentation (WSSS) method that uses image-level annotations to generate pixel-level fibrosis segmentation, reducing the need for fine-grained manual labeling. Additionally, our DiffSeg incorporates a diffusion-based generative model to synthesize HRCT images with different levels of fibrosis from healthy slices, enabling the generation of the fibrosis-injected slices and their paired fibrosis location. Experiments indicate that our method significantly improves the accuracy of pseudo masks generated by existing WSSS methods, greatly reducing the complexity of manual labeling and enhancing the consistency of the generated masks.