Research On Some Issues In Information Hiding Technology For Video And Image

With the development of computer information processing and network transmission capabilities,multimedia information can be easily and efficiently spread in the Internet,and can let people copy and modify arbitrarily.In this trend,the content security of multimedia information has become a common concern in the past two decades.Among the many information security technologies,encryption and information hiding technology has become the main means of multimedia information content privacy protection,tamper detection,copyright management and so on.From the perspective of the embedded domain,information hiding technologies can be divided into spatial domain(or time domain)and transform domain methods.The transform domain method includes the general transform domain(such as DCT)and the encryption domain method which was proposed in recent years.They can also be divided into reversible data hiding and irreversible data hiding according to the reversibility.Based on the different classification of information hiding,this thesis studies the information hiding technology in spatial domain and cryptographic or compressed domain of image and video.This thesis also focuses on the reversibility and capacity-distortion performance in information hiding.The content of this thesis can be divided into three parts,namely,reversible data hiding in spatial domain,DCT domain watermarking algorithm and its optimization,and reversible watermarking of video in the encryption domain.These three parts correspond to the spatial domain,transform domain,and encryption domain method respectively.The main contributions of this thesis are as follows:1.This thesis proposes a pixel prediction-based histogram shift method in order to increase payload and reduce distortion.Our method is designed based on the modified Warped Distance algorithm and the local gradient of an image,which can increase the prediction accuracy and improve the HS algorithm.Experiments show that the proposed method outperforms the previous counterparts in terms of both the prediction accuracy and the final embedding performance and we can run tradeoff between the payload and the distortion by modifying the embedding level as well.2.With the aim of improving the rate ?distortion performance of the watermarked image,a dynamic programming method based on SSIM(structural similarity index)was proposed.The rate-distortion optimization,watermark embedding and JPEG compressing were conducted commutatively so as to improve the compression performance of the watermarked image.Meanwhile,the payload and the robustness was not affected by the optimization procedure through the experimental data.3.We proposed an encrypted domain reversible watermarking scheme and implemented an authenticated encryption system.This scheme,based on the H.265/HEVC video coding standard,separates the reversible watermarking embedding from the encrypting process in the DCT domain so that the Message Authentication Code(MAC)can be embedded into the encrypted video stream.Thereby,the content security and the content authentication can be achieved simultaneously.Compared with other schemes in recent years,these schemes can improve the information hiding capacity and distortion performance,and implement the reversible information hiding,encryption domain and compressed domain embedding and rate-distortion optimization,and provide some feasible ideas to other relevant studies in the future.