AFSegNet: few-shot 3D ankle-foot bone segmentation via hierarchical feature distillation and multi-scale attention and fusion

• Published in: Computerized medical imaging and graphics Vol. 118; p. 102456
• Main Authors: Huang, Yuan, Holcombe, Sven A., Wang, Stewart C., Tang, Jisi
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2024
• Subjects: 3D deep learning; Bone segmentation; Feature fusion; Hierarchical feature distillation; Multi-level attention; Bone segmentation; 3D deep learning; Hierarchical feature distillation; Feature fusion; Multi-level attention
• ISSN: 0895-6111; 1879-0771; 1879-0771
• DOI: 10.1016/j.compmedimag.2024.102456

Accurate segmentation of ankle and foot bones from CT scans is essential for morphological analysis. Ankle and foot bone segmentation challenges due to the blurred bone boundaries, narrow inter-bone gaps, gaps in the cortical shell, and uneven spongy bone textures. Our study endeavors to create a deep learning framework that harnesses advantages of 3D deep learning and tackles the hurdles in accurately segmenting ankle and foot bones from clinical CT scans. A few-shot framework AFSegNet is proposed considering the computational cost, which comprises three 3D deep-learning networks adhering to the principles of progressing from simple to complex tasks and network structures. Specifically, a shallow network first over-segments the foreground, and along with the foreground ground truth are used to supervise a subsequent network to detect the over-segmented regions, which are overwhelmingly inter-bone gaps. The foreground and inter-bone gap probability map are then input into a network with multi-scale attentions and feature fusion, a loss function combining region-, boundary-, and topology-based terms to get the fine-level bone segmentation. AFSegNet is applied to the 16-class segmentation task utilizing 123 in-house CT scans, which only requires a GPU with 24 GB memory since the three sub-networks can be successively and individually trained. AFSegNet achieves a Dice of 0.953 and average surface distance of 0.207. The ablation study and comparison with two basic state-of-the-art networks indicates the effectiveness of the progressively distilled features, attention and feature fusion modules, and hybrid loss functions, with the mean surface distance error decreased up to 50 %. •Foot and ankle bone segmentation in low-resolution CT scans is challenging.•A few-shot deep learning-based framework with hierarchical sub-networks distills coarse-to-fine features progressively.•The 3D framework manages computational costs via successively and individually training each network.•The distilled features, multi-level attention and feature fusion, and hybrid loss functions facilitate accurate segmentation.