Preprint
Jul 2026
Beyond Random Sampling: Distribution-Aware Alignment for Semi-Supervised Medical Image Segmentation
This work introduces an offline Distribution-Aware Sample Selection strategy and proposes the Memory-guided Copy-Paste module, a highly data-efficient framework driven by distribution alignment that effectively mitigates early-stage pseudo-label noise.
Weihao Yan, Yeqiang Qian, Yichan Dong et al.
· 0 citations