A Survey of Privacy Preserving Methods based on Differential Privacy for Medical Data

Using deep learning and federated learning in the medical domain has facilitated convenience in diagnosing and treating patients. However, the current privacy leakage problem has made patients more cautious about personal privacy security. Privacy security has become an urgent issue. Differential privacy is a cryptography technique that makes it impossible for attackers to identify personal data by adding noise to the training process. In this work, we first provide a complete overview of differential privacy technologies. Then, we will review existing differential privacy works in the medical field. Finally, we analyze the existing problems of differential privacy and propose possible avenues for further research.