A research team of Asan Medical Center (AMC) recently won the second prize in the universal medical image segmentation decathlon challenge at the International Conference on Medical Image Computing & Computer Assisted Intervention (MICCAI 2018).

The universal medical image segmentation decathlon challenge was a high profile event of this conference, attended by around 1,600 specialists in medical image processing, in which participants design and apply AI algorithm for segmentation by considering the characteristics of various organs and diseases. This year’s challenge was so difficult that only 19 teams, out of the 189 teams that registered, survived till the finals.

Professor Kim Namkug of the Department of Convergence Medicine, Ms. Lee A-reum, researcher at the Biomedical Engineering Research Center, and Mr. Park Beom-hee, graduate student of the University of Ulsan College of Medicine formed a team named “MI2RL (Medical Imaging and Intelligent Reality Lab)” for this challenge. For universal image segmentation, the MI2RL team developed the cascade U-Net that was designed to simultaneously learn E-Net, which specifies an organ or disease in medical images, and P-Net, which accurately segments an organ or disease in a specific location. Professor Kim Namkug (photo: far left) said, “We will continue studying to establish our competitiveness in the field of medical image segmentation technology to help medical specialists in the clinical setting and enhance the quality of medical services.”