Adaptive Semi-Supervised Segmentation of Brain Vessels with Ambiguous Labels

TL;DR

This paper introduces an adaptive semi-supervised segmentation method for brain vessels that effectively handles ambiguous labels and small vessels, improving accuracy in clinical imaging applications.

Contribution

It presents a novel adaptive semi-supervised approach with progressive learning, adaptive training, and boundary enhancement for better brain vessel segmentation.

Findings

Outperforms existing methods on 3DRA datasets

Effectively segments small and ambiguous vessels

Achieves high accuracy with partially labeled data

Abstract

Accurate segmentation of brain vessels is crucial for cerebrovascular disease diagnosis and treatment. However, existing methods face challenges in capturing small vessels and handling datasets that are partially or ambiguously annotated. In this paper, we propose an adaptive semi-supervised approach to address these challenges. Our approach incorporates innovative techniques including progressive semi-supervised learning, adaptative training strategy, and boundary enhancement. Experimental results on 3DRA datasets demonstrate the superiority of our method in terms of mesh-based segmentation metrics. By leveraging the partially and ambiguously labeled data, which only annotates the main vessels, our method achieves impressive segmentation performance on mislabeled fine vessels, showcasing its potential for clinical applications.