Research on Liver Vessel Image Segmentation Based on DeepLabV3+Model

Liver vessel image segmentation plays an important role in the diagnosis, treatment and operation of liver disease. In this paper, a method of liver vessel image segmentation based on deep learning is proposed. Based on DeepLabV3+ architecture, it integrated ResNet-50 encoder, space pyramid pool module and decoder. Among them, the ResNet-50 encoder enhances the feature extraction capability of the network through the deep residual learning framework. The spatial pyramid pool module captures the context information with multi-scale void-convolution under different void-rates, while the decoder performs upsampling and feature fusion to combine the high-level semantic information captured in the coding stage with the original spatial information. This ensures accurate recovery of details and clear definition of segmentation edges, and gives accurate segmentation on complex hepatic vascular datasets. The experimental results show that the proposed image segmentation method has good feasibility and generalization ability.