A Medical image segmentation model with auto-dynamic convolution and location attention mechanism

Open Webpage

Yuenan Wang, Hua Wang, Fan Zhang

Published: 01 Jan 2025, Last Modified: 19 Jun 2025Comput. Methods Programs Biomed. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Learning different concerns in all four dimensions of nuclear space.•Deconvolution is used for scale adaptation to improve accuracy.•Location Attention can obtain cross-channel information.•Internal Scaling layer can reduce redundancy and improve model efficiency.