Application Of Homomorphic Encryption Algorithm In Encrypted Images

At present,network technology is developing rapidly,but the network security situation becomes increasingly worse.According to the development of the cloud computing platform,people can store and calculate a large amount of data.In the rapid development of big data and the popularization of cloud computing,how to effectively realize the safe use of big data and user privacy protection has become a difficult point in the field of network security research in the future.In order to avoid these problems,the security conditions and technical problems in the process of image data processing are studied and analyzed.Some solutions to protect the image data of the outsourced image while processing and applying the image are proposed.In this dissertation,a comprehensive and detailed research and analysis of the feature extraction of encrypted images based on RLWE homomorphic encryption algorithm is carried out.The research content and achievements of this dissertation are mainly as follows:1.Comparison,optimization and implementation of homomorphic encryption algorithms.In order to cope with the increasing demand of large amount of ciphertext data processing,this dissertation implements,detects and optimizes different homomorphic encryption algorithms.In this dissertation,a variety of RLWE-based full homomorphic encryption schemes are implemented,tested and optimized.These schemes are implemented by using SIMD's Batch(homogeneous)homomorphic calculation method.An improved homomorphic encryption algorithm is designed to improve the security of homomorphic encryption.Ubuntu 12.04 Server LTS system was used as the experimental platform operating system.Through experiments,and based on the obtained results,the completion efficiency of key generation,encryption and decryption time,homomorphic addition operation and homomorphic multiplication operation in these schemes were analyzed.With the results obtained,it can provide efficient,safe and reliable guarantee for the application development of the homomorphic encryption method in practice.2.Research on the application of homomorphic encryption algorithm in encrypted images.It is required to process image ciphertext data under the condition of protecting user privacy and image data security.This dissertation presents an improved homomorphic encrypted image feature detection scheme based on the RLWE problem.Based on SIFT algorithm on the ciphertext domain,this scheme can implement image feature detection.After image feature detection is implemented,feature point matching of the encrypted image is completed using an exhaustive search method.This dissertation has completed high-efficiency ciphertext image feature point detection and high-accuracy image feature point matching,under the premise of protecting the confidentiality of user images and user privacy.