MICCAI 2020 Daily - Wednesday

3 Yingying Zhu 11 have to label lots of images. I wanted to develop something smart and efficient. This idea came to me and I did a lot of research and experiments by myself. I found that patch-based image translation models could solve the problem. My collaborators in NIH are still using my model and it works well .” Thinking about next steps for this work, Yingying is interested in medical image translation across different modalities. The idea for this work came when Yingying was working on a project at the National Institutes of Health (NIH) . Her collaborators wanted a plaque detector to calculate the calcified score of their patients. They only provided a few images in the pre-contrast domain and wanted the plaque detector working for both pre-contrast and post-contrast. “They threw out this problem to me and I was thinking, how can I do this?” Yingying recalls . “I did not want to DAILY Wednesday