CRIS: Collaborative Refinement Integrated with Segmentation for Polyp Segmentation

TL;DR

This paper introduces CRIS, a novel collaborative training strategy that integrates mask refinement with segmentation, significantly improving polyp detection accuracy in medical images, especially under limited data conditions.

Contribution

The paper presents a new collaborative refinement approach that outperforms existing methods in polyp segmentation by integrating mask refinement with segmentation models.

Findings

Superior performance on benchmark datasets

Effective across various segmentation architectures

Enhances polyp detection accuracy

Abstract

Accurate detection of colorectal cancer and early prevention heavily rely on precise polyp identification during gastrointestinal colonoscopy. Due to limited data, many current state-of-the-art deep learning methods for polyp segmentation often rely on post-processing of masks to reduce noise and enhance results. In this study, we propose an approach that integrates mask refinement and binary semantic segmentation, leveraging a novel collaborative training strategy that surpasses current widely-used refinement strategies. We demonstrate the superiority of our approach through comprehensive evaluation on established benchmark datasets and its successful application across various medical image segmentation architectures.