| With the development of cloud computing,more and more people prefer to outsource images to cloud service for saving storage and uploading or downloading images whenever and wherever.There is some sensitive information(such as medical and personal information)in images.To protect the privacy of the user information,the image needs to be encrypted before being uploaded to the cloud.The traditional image encryption technology can hide the privacy information of the original image well,ensuring the image security.But it doesn't support similarity search,which limits the search and use of images.The image encryption supporting image similarity search needs to realize the retrieval of the encrypted image under the premise of ensuring image security.How to balance the relationship between image security and the accuracy of similarity search on encrypted images is the main problem that needs to be solved for image encryption supporting image similarity search.So far,researchers have less research on this issue,and there are some problems,such as low accuracy of similarity search and the lack of image security,in these algorithms.In order to ensure image security and the accuracy of similarity search,this paper proposes a reversible cellular automata image encryption for similarity search.The method mainly includes the following works:(1)Define the similarity search on encrypted images,and analyze its feasibility when implementing pixel-based deterministic encryption.According to this analysis,combined with reversible cellular automata and deterministic encryption,this paper introduces a deterministic encryption based on reversible cellular automata(DERCA).The method supports pixel-granularity,pixel-set-granularity or image-granularity as input,and outputs deterministic encryption results at the corresponding granularity level.(2)According to DERCA,this paper proposes a reversible cellular automata image encryption method which supports similarity search.The method includes two steps,pixel replacement and pixel position scrambling.In the pixel replacement process,pixels are replaced at two levels of granularity by using pixel-based encryption and pixel-set-based encryption.Pixel-based encryption reflects the partial distribution characteristics of the original color histogram,which ensures the accuracy of similarity search on encrypted images.Pixel-set-based encryption plays a role in balancing the histogram of encrypted image,improving image security.Pixel-base encryption and pixel-set-based encryption well solve the contradiction between practicality and security.At the stage of pixel position scrambling,firstly block the image after pixel replacement,and then scramble the pixel position in each block,which improves image security and the accuracy of similarity search at the same time.(3)Based on the proposed image encryption,the method of similarity measurement is given in this paper.The method utilizes the color histogram of each block in the encrypted image to perform similarity measurement,further ensuring the accuracy of similarity search on encrypted images.Experiments on Lena image and Corel-1000 demonstrate that the proposed method not only offers high image security,but also yields superior retrieval accuracy on encrypted images compared with other image encryption algorithms for similarity search. |
