SEAformer: Selective Edge Aggregation transformer for 2D medical image segmentation

Jingwen Li, Jilong Chen, Lei Jiang, Ruoyu Li, Peilun Han, Junlong Cheng

Abstract: Highlights•SEAformer integrates CNNs and Transformers for robust multi-scale feature fusion.•The SEA module adaptively filters noise and focuses on boundary-related features.•A loss function combines edge and content, using MLO to refine segmentation.•Proven efficacy in diverse tasks: Skin, Thyroid, Gland, and COVID-19 segmentation.