Research On Privacy-preserving Distance Calculation Protocol And Its Application

The development of technologies,such as big data and cloud computing,provides people with more and more convenient services.At the same time,the amount of network information processing and interaction is gradually increasing and the risk of malicious use,which may pose security hidden danger to users' information.Therefore,while providing network convenience for users,it is of great significance to consider their information security issues.There are various ways to achieve privacy protection.As an important branch of cryptography,Secure Multi-party Computing(Secure Multi-party Computing,referred to as SMC)has been widely used in scientific calculation,statistical analysis,geometric position determination and other problems.As one of the common problems in Secure Multi-party Computing,privacy-preserving distance calculation is the basis to solve many practical problems.According to the model of problems,the specific problem can be abstracted into different distance calculation problems.Privacy-preserving distance calculation has important research value both in theory and practice,as well as a hot topic research of scholars.This paper focuses on the Euclidean distance protocol and the Mahalanobis distance protocol with privacy protection,and gives specific applications.The works of this paper are as follows:(1)Aiming at the disclosure of sensitive information,such as its center and radius of the circles,in the positional determination of two geometry circles,a privacy-preserving positional determining protocol for two geometry circles is proposed.Considering the existence of negative numbers in our scheme,this paper proposes secure Euclidean distance computation protocol by dividing the plaintext space of Paillier into two equal length intervals to achieve the correct mapping of negative decrypted result in the plaintext space.In addition,based on this protocol we design a privacy-preserving positional determining protocol for two geometry circles which improves the computational efficiency of both parties without divulging the sensitive information such as the radius and center of the two circles.Finally,this paper gives the detailed correctness analysis and security analysis,as well as actual performance test.The result shows that,on the one hand,our scheme can be used to determine the position of inner tangency and inclusion of two circles besides separation,tangency and intersection.On the other hand,our protocol is applicable to the above five position relationships and is low computational cost when two circles are close to each other and far away from each other.Further,we extend the above protocol,then proposes an application which is the method of privacy-preserving close contact identification.(2)Aiming at the data security issues of users in disease diagnosis service,a disease diagnosis scheme based on secure Mahalanobis distance evaluation model is designed.This paper utilizes the homomorphic re-encryption scheme(Homomorphic Re-Encryption Scheme,referred to as HRES)to realize secure computation over the outsourced medical data.Further,this paper provides an extended secure multiplication algorithm to solve the multiplication problem of two encrypted data,and a minimum value comparison algorithm over ciphertext for comparing encrypted Mahalanobis distance.Finally,this paper presents the specific correctness and security analysis.In the performance analysis,its accuracy of our scheme can be verified in real data sheets,and the computational complexity of our scheme is analyzed with the simulated data sheets by simulating the execution of our scheme.The result demonstrates that our proposed scheme has high diagnostic accuracy and low computational cost,which is suitable for practical medical diagnosis services.This paper researches on the problem of privacy-preserving distance calculation,including Euclidean distance and Mahalanobis distance.Then a privacy-preserving positional determining protocol for two geometry circles and a disease diagnosis scheme based on secure Mahalanobis distance evaluation model are proposed respectively.The homomorphic encryption algorithm is applied to realize the secure calculation of data in above two schemes.The actual performance analysis indicates that our schemes has the advantages of low communication overhead and high computational efficiency,and has practical value in the application of close contact identification and disease diagnosis of medical services.