Stimulus-guided adaptive transformer network for retinal blood vessel segmentation in fundus images

Ji Lin, Xingru Huang, Huiyu Zhou, Yaqi Wang, Qianni Zhang

Abstract: Highlights•Inspired by the visual cortex responses to different stimulus types.•A Stimulus-Guided Adaptive (SGA) Module to produce local–global compound features.•A SGA Pooling Transformer to produce context-rich embedding with noise suppression.•A SGA Feature Fusion module to adaptively adjust receptive field based on the task.•An integrated framework for automatic retinal blood vessels segmentation.