Computer Vision News - December 2018

16 Challenge: CHAOS Computer Vision News by M. Alper Selver Clinical Motivation: The segmentation of abdominal organs has critical importance for several clinical procedures such as pre-evaluation of liver for living donor based transplantation surgery or detailed analysis of abdominal organs and vascular tree for correct positioning of a graft onto an aortic aneurysm. This motivates an ongoing research to achieve better segmentation results and requires overcoming countless challenges originating from both highly flexible anatomical properties of abdomen and limitations of imaging modalities. History and Background: In 2007, SLIVER challenge provided a comparative study of a range of algorithms for liver segmentation from CT and reported a snapshot of the methods that were popular for medical image analysis. In 2015, the VISCERAL Anatomy challenge has brought significant contributions to the field. Since then, machine learning based automatic strategies, especially deep learning through convolutional neural networks , introduced significant novelties and improvements to medical image segmentation. We want to observe the contributions of these latest developments through CHAOS. Therefore, while developing and training the algorithms, it is allowed to include other datasets from other sources such as SLIVER or VISCERAL. Besides using other data sets, utilization of new approaches such as The CHAOS - Combined (CT-MR) Healthy Abdominal Organ Segmentation Challenge has two aims: segmentation of liver from CT and segmentation of four abdominal organs. We have asked main organizer M. Alper Selver to give our readers an exclusive overview of this challenge and we are grateful for this contribution. Challenge A. Emre Kavur (holding a 3D printed liver) andM. Alper Selver at the 11th TurkishMedical Informatics Congress (November 2018, Ankara Turkey). 3D visualization of the ground truth and outcomes of different segmentationmethods