Medical Image Segmentation with Graph Reasoning

Abstract

This paper introduces a novel Synergistic Edge-Guided Graph Reasoning Network (SEGRNet) designed to address the limitations of traditional Convolutional Neural Networks (CNNs) in medical image segmentation, particularly in capturing global information and modeling complex topological relationships. Existing CNNbased methods, such as U-Net and its variants, suffer from limited receptive fields, hindering their ability to capture comprehensive global context, especially in structurally complex biological tissues.