Improving cross-domain generalizability of medical image segmentation using uncertainty and shape-aware continual test-time domain adaptation.

Continual test-time adaptation (CTTA) aims to continuously adapt a source-trained model to a target domain with minimal performance loss while assuming no access to the source data. Typically, source models are trained with empirical risk minimization (ERM) and assumed to perform reasonably on the target domain to allow for further adaptation. However, ERM-trained models often fail to perform adequately on a severely drifted target domain, resulting in unsatisfactory adaptation results.

Three-dimensional architecture of the human subscapularis muscle in vivo.

Detailed analysis of skeletal muscle architecture provides insights into skeletal muscle function. To date, measurements of the human subscapularis architecture have been limited to cadaveric measurements. In this study we demonstrate the feasibility of using anatomically constrained fibre tractography to reconstruct and quantify the 3D architecture of the human subscapularis muscle, and provide the first quantitative measurements of the architecture of the human subscapularis muscle in vivo.

Three-dimensional architecture and moment arms of human rotator cuff muscles in vivo: Interindividual, intermuscular, and intramuscular variations.

The human rotator cuff consists of four muscles, each with a complex, multipennate architecture. Despite the functional and clinical importance, the architecture of the human rotator cuff has yet to be clearly described in humans in vivo. The purpose of this study was to investigate the intramuscular, intermuscular, and interindividual variations in architecture and moment arms of the human rotator cuff.

Hybrid dual mean-teacher network with double-uncertainty guidance for semi-supervised segmentation of magnetic resonance images.

Semi-supervised learning has made significant progress in medical image segmentation. However, existing methods primarily utilize information from a single dimensionality, resulting in sub-optimal performance on challenging magnetic resonance imaging (MRI) data with multiple segmentation objects and anisotropic resolution. To address this issue, we present a Hybrid Dual Mean-Teacher (HD-Teacher) model with hybrid, semi-supervised, and multi-task learning to achieve effective semi-supervised segmentation.