Segmentation Features for CT Scans: A Taxonomy
1Youssef Ouassit, Mohamed Azzouazi and Mohammed Yassine EL Ghoumari
Image segmentation is a crucial task in medical imaging applications, segmentation can aid in several medical acts such as planning therapy radiation, automatic labeling of anatomical structures, lesion detection, surgical intervention, virtual surgery simulation, intra-surgery navigation, etc. Despite works done in imaging segmentation it stays challenging because of problems linked to image acquisition conditions and artifacts such as low contrast images, similar intensities with adjacent objects of interests, noise, etc. In the last decade a big variety of algorithms was proposed for this aim. A widely recent used method consists of using artificial intelligent to achieve the segmentation task based on present labeled images. In this paper we review the relevant proposed approaches in medical imaging segmentation, with a focus on the methods based on AI and specially the deep learning methods, we summarize the accurate algorithms in a taxonomy followed by a comparison discussion. Finally, we present the new researches directions that aim at overcoming current limitations in segmentation task.
Segmentation, Medical Image, CT Scans, Image Features, Image Representation, Deep Learning.