Research On Video Encryption Scheme Based On H.265

As an important carrier of information sharing in modern society and with the rapid development of information and communication technology,video has become more and more widely used.In order to solve the challenges of transmission and storage of video signals with huge data,the video compression technology has been updated,and the H.265/HEVC(High Efficiency Video Coding)standard with excellent compression performance will gradually become more and more popular.The openness of the network makes it easier to steal video information.At the same time,with the deepening of people's copyright awareness,the encryption of video signals has become an important research content.Therefore,the research of video encryption scheme combined with H.265 / HEVC standard is as follows:(1)Based on H.265/HEVC,a full-frame encryption scheme combining transform coefficients scrambling and syntax element encryption is proposed.According to the size of the Transform Unit(TU),different scrambling schemes are used.Arnold transform is used to scramble TUs with the size of 4×4,for other sizes of TUs,a shift encryptor initialized according to the approximate distributed law of the DC coefficients of TUs is designed to shifted encrypt the DC coefficients.At the same time,the logistic chaotic sequence is used to encrypt some of the syntax elements on which bypass encoding is used during the entropy encoding process.The experimental results show that the scheme has high security,the visibility of the encrypted test video sequences are extremely low,the peak signal-to-noise ratio(PSNR)is less than 13 d B,and the structural similarity(SSIM)is less than0.55,the compression ratio is only reduced by 1.126%,and it only brings 0.17% increase rate in encoding time.(2)For the purpose of protecting the information of moving objects,the inter-frame difference method is used to detect moving objects,and then the inter-frame defference image is binarized,corroded and expanded to obtain a “foreground binary image”.Counting the value of the pixels at the position corresponding to each coding unit(CU)of the current frame in it,and the CU whose“foregroud” pixel ratio exceeds the threshold is defined as the CU to be encrypted.Then,the threecomponent of YCb Cr and syntax elements in these CUs are encrypted,which can protect the information of the moving objects well.The experimental results show that the more number and the more intense of moving objects included in the video,the better the encryption effect it has,at the same time,the PSNR and SSIM after encrption will decrease with the increase of the frame number,it can satisfy the need of perceptual encryption to attract users to consume by part of the video contents.(3)Sampling of video frames is implemented by using its feature of a large amount of timedomain redundancy,and “binary saliency image” is got after saliency detection,binarization,corrosion and expansion.Different classification levels are set by counting the encoding file sizes of different sequences with different Tiles evenly divided,then the optimal division scheme is selected according to certain rules and the encryption of syntax elements is executed in part of the Tiles,the difference between the “binary saliency images” of adjacent sampling frames is used to dynamically change the division and encryption scheme.The experimental results show that the scheme ensures a good encryption effect on the target of interest,and also has little impact on compression ratio,PSNR and SSIM are at a lower level,at the same time it form a good complement with the second scheme.