Search papers, labs, and topics across Lattice.
This paper introduces RegAL, a novel framework that unifies active learning (AL) and semi-supervised learning (SSL) for medical image segmentation in scenarios with limited annotated data. By employing a shared topology-aware Pareto optimization, RegAL simultaneously selects cases for annotation and utilizes unlabeled data, addressing the challenges of cold-start conditions. The framework demonstrates superior performance over state-of-the-art methods across multiple datasets, achieving stability and improved segmentation metrics even with minimal labeled volumes.
RegAL outperforms traditional methods by effectively integrating active and semi-supervised learning, achieving robust segmentation with minimal labeled data.
In practical settings, medical image segmentation models are often developed with limited annotated data rather than fully labeled datasets. Training frequently begins in ultra-low labeled regimes where only a small number of volumes are annotated. In such scenarios, practitioners must simultaneously decide which cases to annotate and how to best use the remaining unlabeled data. Although active learning (AL) and semi-supervised learning (SSL) both target annotation scarcity, they are typically designed and optimized independently, resulting in objective mismatch and unstable training during early-stage"cold start"conditions. We propose RegAL, a unified active semi-supervised framework governed by a shared topology-aware Pareto optimization that couples sample acquisition with unlabeled data utilization. RegAL evaluates images along three complementary axes, voxel-wise uncertainty, feature diversity, and a novel topological consistency metric, to select anatomically informative edge cases for annotation. On the other hand, the same criteria are used to identify geometrically stable atlas candidates for diffeomorphic registration-guided augmentation to train a self-supervised Mean Teacher segmentation network. Across BraTS 2021, dHCP, and ProstateX, RegAL remains stable with few labeled volumes and consistently outperforms state-of-the-art AL, SSL, and active semi-supervised baselines across Dice and boundary-distance (ASD, HD95) metrics under extreme annotation scarcity.
