Research On Medical Data Privacy Protection Application Based On Homomorphic Encryption

In recent years,the scale of medical data is constantly growing with the rapid continuous process of medical informatization construction.By analyzing and researching on the medical data,we may get the relations among those data to provide a better medical service and health care for the patients.Commonly,the medical data,however,contains a lot of private information.It has become an important issue on how to protect the private information by limiting medical data analysis.This thesis firstly proposes a medical data privacy protection scheme based on homomorphic encryption by consisting of data collection terminal,edge server,cloud computing server and client.In this scheme,the data collection terminal collects medical data and encrypts it by using a public key encryption scheme;the edge server encrypts the data from the public key encryption algorithm to the homomorphic encryption algorithm;the major task of the cloud computing server is the storage and processing of the encrypted data.The scheme not only analyzes and processes the medical data,but also protects the privacy in the stages of transmission,storage and processing.According to the above mentioned scheme,this thesis also proposes a classification algorithm in the ciphertext domain.The algorithm is introduced based on the homomorphism of the homomorphic encryption algorithm,by combining with a numerical comparator,to take a k-nearest neighbor classification for the encrypted data.Furthermore,a statistical scheme under the ciphertext domain is also proposed in this thesis for meeting the requirements of daily data analysis.This scheme is used to calculate the average value,the range,the extreme value and other statistic of the encrypted data without sensitive data leakage during such processing.The simulation result shows that the processing efficiency of the homomorphic encryption algorithm used in our proposed scheme is available to fulfill the requirement of the daily medical data analysis.Under the premise of ensuring the privacy of the user's medical data,the proposed schemes can correctly calculate the average value and extreme value of the ciphertext data,and can classify the ciphertext data by k-nearest neighbor,to reduce the pressure of computing and storage from clients.