A review of research on federated learning in the field of medical image processing

Le résumé fourni par la source

In the field of medical image processing, data sharing is difficult to achieve due to privacy protection issues, and there are also risks such as high communication overhead and privacy leakage in multi center electronic case data. Federated learning, as a method that allows multiple participants to jointly train a shared machine learning model without sharing raw data, can technically break down data silos and has been applied in multiple industries. The problem of data silos has long led to the generalization issue of machine learning in the field of medical image processing. The emergence of federated learning has brought hope to solve the problem of data sharing in medical image processing. In recent years, many researchers have conducted extensive research on the problems in the field of medical image processing using federated learning. However, there is currently a lack of inductive analysis of these studies, which is not conducive to the development of subsequent research. This article provides a brief introduction to federated learning and lists some of its applications in the field of medical image processing. It summarizes the four directions in which federated learning is applied in the medical field: disease classification, disease prediction, medical image synthesis, and medical image segmentation. Finally, the challenges of federated learning were summarized and some solutions were proposed, hoping to provide some assistance for future research.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.