Scribble-Supervised Medical Image Segmentation with Dynamic Teacher Switching and Hierarchical Consistency

TL;DR

This paper introduces SDT-Net, a novel framework for scribble-supervised medical image segmentation that employs dynamic teacher switching and hierarchical consistency to improve segmentation accuracy despite sparse annotations.

Contribution

The paper proposes a dual-teacher, single-student framework with adaptive teacher selection and hierarchical consistency mechanisms to enhance scribble-supervised segmentation.

Findings

Achieves state-of-the-art results on ACDC and MSCMRseg datasets.

Produces more accurate and anatomically plausible segmentations.

Effectively handles annotation sparsity and ambiguity.

Abstract

Scribble-supervised methods have emerged to mitigate the prohibitive annotation burden in medical image segmentation. However, the inherent sparsity of these annotations introduces significant ambiguity, which results in noisy pseudo-label propagation and hinders the learning of robust anatomical boundaries. To address this challenge, we propose SDT-Net, a novel dual-teacher, single-student framework designed to maximize supervision quality from these weak signals. Our method features a Dynamic Teacher Switching (DTS) module to adaptively select the most reliable teacher. This selected teacher then guides the student via two synergistic mechanisms: high-confidence pseudo-labels, refined by a Pick Reliable Pixels (PRP) mechanism, and multi-level feature alignment, enforced by a Hierarchical Consistency (HiCo) module. Extensive experiments on the ACDC and MSCMRseg datasets demonstrate that SDT-Net achieves state-of-the-art performance, producing more accurate and anatomically plausible segmentation.