Discriminative dictionary learning for abdominal multi-organ segmentation

• Published in: Medical image analysis Vol. 23; no. 1; pp. 92 - 104
• Main Authors: Tong, Tong, Wolz, Robin, Wang, Zehan, Gao, Qinquan, Misawa, Kazunari, Fujiwara, Michitaka, Mori, Kensaku, Hajnal, Joseph V., Rueckert, Daniel
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.07.2015
• Subjects: Abdominal multi-organ segmentation; Algorithms; Discriminative dictionary learning; Humans; Kidney - diagnostic imaging; Liver - diagnostic imaging; Local atlas selection; Pancreas - diagnostic imaging; Patch based; Radiographic Image Interpretation, Computer-Assisted - methods; Radiography, Abdominal - methods; Spleen - diagnostic imaging; Tomography, X-Ray Computed - methods; Local atlas selection; Abdominal multi-organ segmentation; Discriminative dictionary learning; Patch based
• ISSN: 1361-8415; 1361-8423; 1361-8423
• DOI: 10.1016/j.media.2015.04.015

•Discriminative dictionary learning is applied to multi-organ segmentation.•A local voxel-wise atlas selection is proposed.•A multi-resolution strategy is developed for gaining computational efficiency.•Validation is carried out on 150 abdominal CT images.•A comparison between different atlas selection strategies. The framework of the multi-resolution segmentation process. The proposed DDLS approach is performed to generate probabilistic atlas for each organ, which propagates across resolutions. The segmentation mask at current resolution is limited to the voxels with uncertain segmentations at the previous resolution. The final segmentation is achieved by using the graph-cuts algorithm in the native space. [Display omitted] An automated segmentation method is presented for multi-organ segmentation in abdominal CT images. Dictionary learning and sparse coding techniques are used in the proposed method to generate target specific priors for segmentation. The method simultaneously learns dictionaries which have reconstructive power and classifiers which have discriminative ability from a set of selected atlases. Based on the learnt dictionaries and classifiers, probabilistic atlases are then generated to provide priors for the segmentation of unseen target images. The final segmentation is obtained by applying a post-processing step based on a graph-cuts method. In addition, this paper proposes a voxel-wise local atlas selection strategy to deal with high inter-subject variation in abdominal CT images. The segmentation performance of the proposed method with different atlas selection strategies are also compared. Our proposed method has been evaluated on a database of 150 abdominal CT images and achieves a promising segmentation performance with Dice overlap values of 94.9%, 93.6%, 71.1%, and 92.5% for liver, kidneys, pancreas, and spleen, respectively.