Semi-supervised Strong-Teacher Consistency Learning for few-shot cardiac MRI image segmentation

Yuting Qiu, James Meng, Baihua Li

Abstract: Highlights•A multi-teacher approach mitigates unreliable guidance and diversifies learning.•Perturb the feature space via Virtual Adversarial Training to improve generalization.•A novel hybrid loss to address over-regularization and balanced consistency learning.•Comprehensive evaluation and ablation studies on two multi-class benchmark datasets.

External IDs:dblp:journals/cmpb/QiuML25