Computer Vision News - March 2017

Every month, Computer Vision News reviews a challenge related to our field. If you do not take part in challenges, but are interested to know the new methods proposed by the scientific community to solve them, this section is for you. The challenge of this month was launched to provide the endoscopists with tools to identify suspicious areas related to gastroenterological diseases: the ISBI Challenge on Analysis of Images to Detect Abnormalities in Endoscopy (AIDA-E) is divided in 3 sub-challenges; we are going to review the 3 rd one, Gastric chromoendoscopy images in cancer surveillance . The website of the challenge is here . The increase in occurrence of gastroenterological diseases has generated bottlenecks commonly found in endoscopic surveillance , challenging the available resources for monitoring at risk patients and the possibility of early detection and accurate staging of lesions and tissue alterations. The need to relieve the clinicians and health system of part of the bottlenecks and to reduce the number of unnecessary biopsy taken from the patients, has asked for an ever increasing effort from the imaging community to provide the endoscopists with tools able to identify suspicious areas and to accurately assess the patient’s status and needs. Unneeded anxiety stemming from an excessive use of biopsy is another reason to improve the accuracy of screening at an early stage, which can be obtained by developing an effective Mucosa classification model in gastric chromoendoscopy for cancer testing . Participants to this gastric cancer detection challenge were given labeled data, belonging to either groups I (Normal), II or III (Abnormal). Gold- standard final annotations were performed independently by Dr. Dinis- Ribeiro and Dr. Miguel Areia , using a relevant taxonomy according to which chromoendoscopy images are classified into their respective classes based on color, shape and regularity of pit patterns. Participants to the challenge classified the manually segmented region of each image in the dataset as either ‘Normal’ or ‘Abnormal’. The winner introduced a Superpixel-based classification of gastric chromoendoscopy images : Davide Boschetto ’s paper is here and it has being presented at SPIE Medical Imaging in Orlando, FL only two weeks ago. Challenge Analysis of Images to Detect Abnormalities in Endoscopy (AIDA-E): Part 3 Chromoendoscopy Computer Vision News Challenge 31