Design Of Ciphertext Image Retrieval System In Cloud Environment

With the popularity of handheld imaging devices,image data has increased with the explosive trend.However,it is difficult for data owners to store massive image data cheaply,manage effectively and search conveniently.With the development of cloud computing technology,cloud server has the advantages of high-cost performance,centralized management and online upgrade,which solves the problem of users'management of massive image data.However,some cloud service providers leaked or lost users' data in recent years.It made users doubt whether cloud services can guarantee data security.Therefore,how to ensure the security of data has become the primary problem.Usually,users need to encrypt private images before uploading to the cloud server in order to prevent privacy disclosure.However,the relevance of data and features disappeared,which leads to users can't use the whole image and can't meet the actual needs.Therefore,it is necessary to put a ciphertext image retrieval scheme under the cloud computing environment.This paper mainly studies the ciphertext image retrieval technology in the cloud computing environment,which is how to make the image owners can encrypt and store their private images and enable authorized image users to find the images from the massive images quickly.The main contributions of this paper are as follows:(1)A ciphertext image retrieval method based on SimHash in the cloud environment is proposed.First,the data owner extracts the local features of the image and cluster them through K-means so that image information can be represented by visual word codebook.Then,the codebook is hashed into the corresponding fingerprint.Finally,the searchable encryption feature algorithm based on SimHash can directly generate image encryption feature vectors and upload them to the cloud server without building an index.The query user constructs the trap gate for the query image through the same operation and uploads it to the cloud server,which performs similarity retrieval according to the trap gate and image encryption feature vector.The experimental results show that the retrieval accuracy is improved and the memory consumption is reduced.(2)The existing retrieval schemes leave the pre-processing operations such as feature extraction to the data owners,which increases the burden on users and the risk of information leakage.To solve this problem,a searchable image encryption algorithm is proposed in this paper.First,the image is encrypted by Zigzag scanning and image sequence encryption algorithm,and then the feature extractor is uploaded to the cloud.The cloud extracts the encrypted image features through the feature extractor and'generates image vectors for similarity retrieval,to realize the similarity retrieval of encrypted images.(3)We use Matlab to realize the local construction of the encrypted image retrieval system including the system framework design and system interface design.We finally realize the encrypted image retrieval from the data owner module,query user module,query result display module.