Dual-scale enhanced and cross-generative consistency learning for semi-supervised medical image segmentation

Yunqi Gu, Tao Zhou, Yizhe Zhang, Yi Zhou, Kelei He, Chen Gong, Huazhu Fu

Abstract: Highlights•A novel DEC-Seg framework is proposed for semi-supervised medical image segmentation.•We propose a CFA module to fuse features from two adjacent layers.•A DCF module fully exploits scale information to produce more accurate segmentation maps.•Extensive experimental results show the effectiveness of the proposed method.