SCCNet: Self-correction boundary preservation with a dynamic class prior filter for high-variability ultrasound image segmentation. (March 2023)

• Record Type: Journal Article
• Authors: Gong, Yuxin; Zhu, Haogang; Li, Jixing; Yang, Jingchun; Cheng, Jian; Chang, Ying; Bai, Xiao; Ji, Xunming
• Is Part Of: Computerized medical imaging and graphics. Volume 104(2023)
• Publication Date: 2023-03
• Subjects: dynamic Boundary preservation -- Class context -- High variability -- Cascade -- Ultrasound image; Diagnostic imaging -- Periodicals; Imaging systems in medicine -- Periodicals; Diagnosis, Radioscopic -- Data processing -- Periodicals; Diagnostic Imaging -- Periodicals; Imagerie pour le diagnostic -- Périodiques; Diagnostic imaging; Periodicals; Electronic journals; Electronic journals; 616.0754
• DOI: 10.1016/j.compmedimag.2023.102183
• Languages: English
• ISSNs: 0895-6111

Abstract

Abstract: The highly ambiguous nature of boundaries and similar objects is difficult to address in some ultrasound image segmentation tasks, such as neck muscle segmentation, leading to unsatisfactory performance. Thus, this paper proposes a two-stage network called SCCNet (self-correction context network) using a self-correction boundary preservation module and class-context filter to alleviate these problems. The proposed self-correction boundary preservation module uses a dynamic key boundary point (KBP) map to increase the capability of iteratively discriminating ambiguous boundary points segments, and the predicted segmentation map from one stage is used to obtain a dynamic class prior filter to improve the segmentation performance at Stage 2. Finally, three datasets, Neck Muscle, CAMUS and Thyroid, are used to demonstrate that our proposed SCCNet outperforms other state-of-the art methods, such as BPBnet, DSNnet, and RAGCnet. Our proposed network shows at least a 1.2–3.7% improvement on the three datasets, Neck Muscle, Thyroid, and CAMUS. The source code is available at https://github.com/lijixing0425/SCCNet. Highlights: First, we proposed a new Self-Correction boundary preservation module, named DBPB, which is an iterative training strategy proved to have effect on decreasing ambiguous boundary problem for Ultrasound image segmentation with high-variability. Second, we propose a new Dynamic Class Prior Filter(DCPF) module that it incorporates a class-level context prior to construct the dynamic multiscale filter. The proposed DCPF outperforms DCM (Dynamic Convolutional Module), and other popular context modules for ultrasound image segmentation with high-variability, proved by our ablation experiments. Third, we demonstrate that the proposed SCCNet, considering boundary and context modules outperforms the-state-of-the-art networks such as BPBnet, RAGCnet which only considering one of context and boundary modules. Meanwhile, SCCNet outperforms LFB-Neton CAMUS dataset, further demonstrating that our method consistently outperforms the state-of-the-art methods.