Research On Privacy Protection And Search Technology Of Outsourced Data In Cloud Storag

With the advent of the era of big data,the data of users(individuals or enterprises)increases dramatically.Due to the limitation of local storage resources,users can only outsource their data to the cloud platform.However,the privacy and searchability of outsourced data becomes the main barrier hindering the healthy development of cloud storage.In recent years,a lot of studies and solution to solve above issue have been put forward in academia and industry.This thesis mainly focuses on the privacy protection and efficient search of outsourced data in cloud storage.Specifically,it centers on the following two research contents.For one thing,the 0/1-encoding technology used in some searchable encryption schemes is deeply studied and some suggestions on how to use this technique are given to protect the privacy and security of data.For another,on the basis of privacy protection of outsourced medical images,efficient image search is realized.Around these two points,the main contributions of this thesis are as follows.1.Based on a new 0/1-encoding technique used in searchable encryption schemes,this thesis analyzes the existing security problems of 0/1-encoding,and further studies the inherent properties of this encoding method,and measures the degree of data privacy protection of this encoding technique.At the same time,this thesis points out that the 0/1-encoding set should be properly blinded or encrypted to protect the privacy of numerical values when using it for secret numerical comparison.2.Based on the idea of classification,this thesis proposes a privacy-preserving and fast outsourced medical image search scheme.The scheme simply classifies the outsourced images,which greatly improves the efficiency of image search.And our scheme also adopts flexible access control strategy to further realize the privacy protection of outsourced medical images.Considering the high accuracy of medical image search,convolutional neural network is used to extract image features,and the interference of random numbers on relevant score of images is reduced to 0.Finally,a large number of experiments show that our proposed scheme has good performance.