An Overview of Abdominal Multi-organ Segmentation
Qiang Li, Hong Song, Lei Chen, Xianqi Meng, Jian Yang, Le Zhang
Abstract
The segmentation of multiple abdominal organs of the human body from images with different modalities is challenging because of the inter-subject variance among abdomens, as well as the complex intra-subject variance among organs. In this paper, the recent methods proposed for abdominal multi-organ segmentation (AMOS) on medical images in the literature are reviewed. The AMOS methods can be categorized into traditional and deep learning-based methods. First, various approaches, techniques, recent advances, and related problems under both segmentation categories are explained. Second, the advantages and disadvantages of these methods are discussed. A summary of some public datasets for AMOS is provided. Finally, AMOS remains an open issue, and the combination of different methods can achieve improved segmentation performance.