RADA: Region-Aware Dual-encoder Auxiliary learning for Barely-supervised Medical Image Segmentation

TL;DR

RADA introduces a region-aware dual-encoder framework leveraging pre-trained visual features and semantic guidance to improve barely-supervised medical image segmentation with sparse annotations.

Contribution

It proposes a novel dual-encoder auxiliary learning pipeline that enhances segmentation accuracy using region-specific features and semantic supervision.

Findings

Achieves state-of-the-art results on LA2018, KiTS19, and LiTS datasets.

Demonstrates robustness and generalization with extremely sparse annotations.

Outperforms existing methods by integrating fine-grained visual features with semantic guidance.

Abstract

Deep learning has greatly advanced medical image segmentation, but its success relies heavily on fully supervised learning, which requires dense annotations that are costly and time-consuming for 3D volumetric scans. Barely-supervised learning reduces annotation burden by using only a few labeled slices per volume. Existing methods typically propagate sparse annotations to unlabeled slices through geometric continuity to generate pseudo-labels, but this strategy lacks semantic understanding, often resulting in low-quality pseudo-labels. Furthermore, medical image segmentation is inherently a pixel-level visual understanding task, where accuracy fundamentally depends on the quality of local, fine-grained visual features. Inspired by this, we propose RADA, a novel Region-Aware Dual-encoder Auxiliary learning pipeline which introduces a dual-encoder framework pre-trained on Alpha-CLIP to extract fine-grained, region-specific visual features from the original images and limited annotations. The framework combines image-level fine-grained visual features with text-level semantic guidance, providing region-aware semantic supervision that bridges image-level semantics and pixel-level segmentation. Integrated into a triple-view training framework, RADA achieves SOTA performance under extremely sparse annotation settings on LA2018, KiTS19 and LiTS, demonstrating robust generalization across diverse datasets.