Boosting knowledge diversity, accuracy, and stability via tri-enhanced distillation for domain continual medical image segmentation

Zhanshi Zhu, Xinghua Ma, Wei Wang, Suyu Dong, Kuanquan Wang, Lianming Wu, Gongning Luo, Guohua Wang, Shuo Li

Abstract: Highlights•Novel framework mitigates catastrophic forgetting.•New augmentation method distills diverse old knowledge.•Adaptive transfer method captures accurate old knowledge.•Global uncertainty-guided fusion stabilizes knowledge fusion.