DGCBG-Net: A dual-branch network with global cross-modal interaction and boundary guidance for tumor segmentation in PET/CT images

Ziwei Zou, Beiji Zou, Xiaoyan Kui, Zhi Chen, Yang Li

Abstract: Highlights•DGCBG-Net aims to address the insufficient cross-modal interaction and ambiguous segmentation of tumor boundaries.•Cross-modal fusion of global features facilitates the features interaction between PET and CT images.•Multi-stage boundary feature extraction is employed to fully explore the potential boundary features of CT images.•The experimental results demonstrate that DGCBG-Net outperforms existing methods, and is competitive to state-of-arts.