A deep learning framework for pancreas segmentation with multi-atlas registration and 3D level-set

• Published in: Medical image analysis Vol. 68; p. 101884
• Main Authors: Zhang, Yue, Wu, Jiong, Liu, Yilong, Chen, Yifan, Chen, Wei, Wu, Ed. X., Li, Chunming, Tang, Xiaoying
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.02.2021; Elsevier BV
• Subjects: Algorithms; Artificial neural networks; Computed tomography; Datasets; Deep learning; Image processing; Image segmentation; Information processing; Level-set; Machine learning; Multi-atlas registration; Neural networks; Pancreas; Pancreas segmentation; Registration; Deep learning; Multi-atlas registration; Level-set; Pancreas segmentation
• ISSN: 1361-8415; 1361-8423
• DOI: 10.1016/j.media.2020.101884

•Segmentation model based on 3D location, contextual, shape and edge information.•Multi-atlas based 3D diffeomorphic registration and fusion.•Joint 2D + 3D deep learning strategy•3D level-set integrating fuzzy c-means. [Display omitted] In this paper, we propose and validate a deep learning framework that incorporates both multi-atlas registration and level-set for segmenting pancreas from CT volume images. The proposed segmentation pipeline consists of three stages, namely coarse, fine, and refine stages. Firstly, a coarse segmentation is obtained through multi-atlas based 3D diffeomorphic registration and fusion. After that, to learn the connection feature, a 3D patch-based convolutional neural network (CNN) and three 2D slice-based CNNs are jointly used to predict a fine segmentation based on a bounding box determined from the coarse segmentation. Finally, a 3D level-set method is used, with the fine segmentation being one of its constraints, to integrate information of the original image and the CNN-derived probability map to achieve a refine segmentation. In other words, we jointly utilize global 3D location information (registration), contextual information (patch-based 3D CNN), shape information (slice-based 2.5D CNN) and edge information (3D level-set) in the proposed framework. These components form our cascaded coarse-fine-refine segmentation framework. We test the proposed framework on three different datasets with varying intensity ranges obtained from different resources, respectively containing 36, 82 and 281 CT volume images. In each dataset, we achieve an average Dice score over 82%, being superior or comparable to other existing state-of-the-art pancreas segmentation algorithms.