SPGNet: A shape-prior guided network for medical image segmentation

Song, Zhengxuan; Liu, Xun; Zhang, Wenhao; Gong, Yongyi; Hao, Tianyong; Zeng, Kun

Abstract

Given the intricacy and variability of anatomical structures in medical images, some methods employ shape priors to constrain segmentation. However, limited by the representational capability of these priors, existing approaches often struggle to capture diverse target structure morphologies. To address this, we propose SPGNet to guide segmentation by fully exploiting category-specific shape knowledge. The key idea is to enable the network to perceive data shape distributions by learning from statistical shape models. We uncover shape relationships via clustering and obtain statistical prior knowledge using principal component analysis. Our dual-path network comprises a segmentation path and a shape prior path that collaboratively discern and harness shape prior distribution to improve segmentation robustness. The shape-prior path further serves to refine shapes iteratively by cropping features from the segmentation path, guiding the segmentation path and directing attention specifically to the edges of shapes which could be most significantly susceptible to segmentation error. We demonstrate superior performance on chest X-ray and breast ultrasound benchmarks.

• Presentation Conference Type: Conference Paper (Published)
• Conference Name: The 33rd International Joint Conference on Artificial Intelligence
• Start Date: Aug 3, 2024
• End Date: Aug 9, 2024
• Acceptance Date: Apr 16, 2024
• Publication Date: Aug 9, 2024
• Deposit Date: May 27, 2024
• Pages: 1263-1271
• Book Title: Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence
• DOI: https://doi.org/10.24963/ijcai.2024/140
• Public URL: https://uwe-repository.worktribe.com/output/12008166