S&D Messenger: Exchanging Semantic and Domain Knowledge for Generic Semi-Supervised Medical Image Segmentation

TL;DR

This paper introduces S&D Messenger, a framework that exchanges semantic and domain knowledge between labeled and unlabeled data, significantly improving semi-supervised medical image segmentation across various tasks.

Contribution

The paper proposes a novel knowledge exchange framework that unifies semi-supervised segmentation, domain generalization, and domain adaptation in medical imaging.

Findings

Achieves +7.5% improvement on six datasets for semi-supervised segmentation.

Enhances performance on domain adaptation and generalization tasks.

Enables a simple pseudo-labeling method to outperform state-of-the-art approaches.

Abstract

Semi-supervised medical image segmentation (SSMIS) has emerged as a promising solution to tackle the challenges of time-consuming manual labeling in the medical field. However, in practical scenarios, there are often domain variations within the datasets, leading to derivative scenarios like semi-supervised medical domain generalization (Semi-MDG) and unsupervised medical domain adaptation (UMDA). In this paper, we aim to develop a generic framework that masters all three tasks. We notice a critical shared challenge across three scenarios: the explicit semantic knowledge for segmentation performance and rich domain knowledge for generalizability exclusively exist in the labeled set and unlabeled set respectively. Such discrepancy hinders existing methods from effectively comprehending both types of knowledge under semi-supervised settings. To tackle this challenge, we develop a Semantic & Domain Knowledge Messenger (S&D Messenger) which facilitates direct knowledge delivery between the labeled and unlabeled set, and thus allowing the model to comprehend both of them in each individual learning flow. Equipped with our S&D Messenger, a naive pseudo-labeling method can achieve huge improvement on six benchmark datasets for SSMIS (+7.5%), UMDA (+5.6%), and Semi-MDG tasks (+1.14%), compared with state-of-the-art methods designed for specific tasks.