RT Journal Article ID 5fab7b697ba44b4d A1 Wu, Jie A1 Kamath, Markad A1 Noseworthy, Michael D. A1 Boylan, Colm A1 Poehlman, Skip T1 Segmentation of Images of Abdominal Organs JF Critical Reviews™ in Biomedical Engineering JO CRB YR 2008 FD 2008-12-01 VO 36 IS 5-6 SP 305 OP 334 K1 medical image processing K1 image segmentation K1 computed tomogra¬phy K1 magnetic resonance imaging K1 abdomen K1 organ recognition K1 pattern recognition K1 deformable model AB Abdominal organ segmentation, which is, the delineation of organ areas in the abdomen, plays an important role in the process of radiological evaluation. Attempts to automate segmentation of abdominal organs will aid radiologists who are required to view thousands of images daily. This review outlines the current state-of-the-art semi-automated and automated methods used to segment abdominal organ regions from computed tomography (CT), magnetic resonance imaging (MEI), and ultrasound images. Segmentation methods generally fall into three categories: pixel based, region based and boundary tracing. While pixel-based methods classify each individual pixel, region-based methods identify regions with similar properties. Boundary tracing is accomplished by a model of the image boundary. This paper evaluates the effectiveness of the above algorithms with an emphasis on their advantages and disadvantages for abdominal organ segmentation. Several evaluation metrics that compare machine-based segmentation with that of an expert (radiologist) are identified and examined. Finally, features based on intensity as well as the texture of a small region around a pixel are explored. This review concludes with a discussion of possible future trends for abdominal organ segmentation. PB Begell House LK https://www.dl.begellhouse.com/journals/4b27cbfc562e21b8,1d0b7cfe02b75d39,5fab7b697ba44b4d.html