Automatic Multi-organ Segmentation Using Learning-Based Segmentation and Level Set Optimization

• DOI: 10.1007/978-3-642-23626-6_42
• Publication Date: 2011-09-20T06:18:46Z
• Authors: Timo Kohlberger; Michal Sofka; Jingdan Zhang; Neil Birkbeck; Jens Wetzl; Jens Kaftan; Jérôme Declerck; S. Kevin Zhou
• Keywords: Models, Anatomic; Principal Component Analysis; Models, Statistical; Databases, Factual; Reproducibility of Results; 02 engineering and technology; Kidney; Pattern Recognition, Automated; 03 medical and health sciences; 0302 clinical medicine; Imaging, Three-Dimensional; Liver; Artificial Intelligence; 0202 electrical engineering, electronic engineering, information engineering; Humans; Learning; Tomography, X-Ray Computed; Lung; Algorithms; Software

We present a novel generic segmentation system for the fully automatic multi-organ segmentation from CT medical images. Thereby we combine the advantages of learning-based approaches on point cloud-based shape representation, such a speed, robustness, point correspondences, with those of PDE-optimization-based level set approaches, such as high accuracy and the straightforward prevention of segment overlaps. In a benchmark on 10-100 annotated datasets for the liver, the lungs, and the kidneys we show that the proposed system yields segmentation accuracies of 1.17-2.89 mm average surface errors. Thereby the level set segmentation (which is initialized by the learning-based segmentations) contributes with an 20%-40% increase in accuracy.