MICCAI 2022 Daily – Wednesday

Federated learning has become a promising solution for collecting data from different hospitals and medical institutions for collaborative training. Most existing works consider supervised learning, which means datasets must be annotated. However, many medical institutions do not have the time or the budget to do the labeling. In this paper, Meirui and the team propose a new solution to this real- world problem, enabling those unlabeled clients to be a part of federated learning . Meirui Jiang is a third-year PhD student at the Chinese University of Hong Kong, majoring in federated learning and medical image analysis under the supervision of Qi Dou. His paper proposes a novel solution to a semi-supervised federated learning problem. He speaks to us ahead of his poster today . Dynamic Bank Learning for Semi-supervised Federated Image Diagnosis with Class Imbalance 10 DAILY MICCAI Wednesday Poster Presentation