Computer Vision News - January 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. This month we have chosen to review the Tumor Proliferation Assessment Challenge , organized around MICCAI 2016, which was held 3 months ago in Athens. TUPAC16 is organized by the Medical Image Analysis Group (IMAG/e) of the Eindhoven University of Technology and by the University Medical Center Utrecht, both in the Netherlands. The website of the challenge, with all its related resources, is here . Background Aggressive tumors, typically growing with high proliferation speed, need to be dealt with aggressive therapy. Quick proliferation in breast cancers is statistically followed by worse outcomes. Accurate assessment of proliferation speed provides a precious indicator to the physician who needs to decide how the patient should be treated. Pathologists assessing tumor proliferation grade cancer by counting mitotic figures in H&E (hematoxylin and eosin) stained histological slide preparations observed under a microscope. Proliferation speed is high when density of mitotic figures correlated is high too. Motivation This task is of key medical importance. However, the process of mitosis counting by pathologists is susceptible to subjectivity. Therefore, much effort has been done in finding accurate methods for an objective counting (see this article on automatic mitosis detection using deep neural networks by RSIP Vision ). Challenge 30 Computer Vision News Challenge TUPAC16 - Tumor Proliferation Assessment Challenge 2016 Examples of mitotic figures, shown by the yellow arrows Accurate assessment of proliferation speed provides a precious indicator of how the patient should be treated