Research On Robust Video Watermarking Algorithms For HEVC Copyright Protection

With the rapid development of information technology and the popularization of high efficiency video coding(HEVC),HEVC video plays an more important role in people's production and life.However,due to the lack of effective copyright protection measures for HEVC videos,the circulation of pirated videos is rampant,and various infringement issues and disputes over rights protection emerge one after another.Therefore,copyright protection for HEVC videos is imminent.Robust video watermarking is an important technical means for resolving video copyright disputes.Two types of robust video watermarking:original-video watermarking and compressed-domain watermarking are researched in this thesis.The main tasks completed are as follows:(1)To solve the problems that the existing original-video watermarking algorithms are not imperceptible enough and difficult to resist HEVC compression attack,a robust video watermarking algorithm based on inter-frame discrete cosine transform(DCT)coefficient difference and just noticeable difference(JND)is proposed.To make the propose algorithm resistant to frame-based attacks,the video frames are grouped and embedded into watermark.Because of the stability of the inter-frame DCT coefficient difference,it is regarded as a robust feature.To embed the watermark,the DCT coefficients of the embedding frames are modified based on JND threshold,making the difference between the average value of the embeddings frame coefficients and the average value of the reference frame coefficients within a preset range,which ensures the proposed algorithm has good robustness in the case of imperceptible distortion.In the extraction stage,the averaging method is used for post-processing to further improve the robustness of the algorithm.Taking part of the JCT-VC standard test sequences as the experimental object,the performance of the proposed algorithm is tested.Experimental results show that compared with the existing DCT-based video watermarking algorithms,the proposed algorithm achieves a higher mean opinion score(MOS),which proves that the algorithm has better subjective imperceptibility.At the same time,the algorithm's robustness to HEVC compression,noise,and frame-based attacks is also better than the existing algorithms,with lower bit error rate(BER)and higher normalized cross correlation(NCC).This algorithm can still extract the watermark even when the attack intensity is high,and has a good copyright certification ability,which can effectively prove the ownership of the video copyright.(2)To solve the problem that the existing watermarking algorithms in HEVC compression domain cannot achieve no intra-frame distortion drift,resistance to recompression and signal processing attacks at the same time,a robust compression-domain watermarking algorithm without intra-frame distortion drift based on prediction residual is proposed.First,after research,it is found that a prediction pixel in HEVC video has a robust relationship with the mean value of its 8-neighboring pixels.Based on this discovery,the multi-coefficient modification method is used to embed the watermark into the intra-frame prediction residual pixel of the luminance transform block(LTB),so that the watermark algorithm can resist recompression attack while avoiding intra-frame distortion drift.The residual-based watermark embedding enables the algorithm to extract the watermark information from merged 8×8 blocks after being attacked,which improves the robustness of the algorithm.In addition,three different thresholds are set according to application requirements to improve the performance of the watermarking algorithm,and further improve the imperceptibility and robustness of the algorithm.Taking part of the JCT-VC standard test sequences as the object to conduct experiments,the results show that the proposed algorithm has higher peak signal-to-noise ratio(PSNR)and MOS compared with the existing algorithms,that is,it has better imperceptibility.In addition,the algorithm has lower BER and higher NCC than existing algorithms in HEVC recompression,noise,and filtering attacks tests,which proves that the proposed algorithm is more robust.At the same time,the proposed algorithm embeds watermarks in the process of video compression and uses less computing resources,which is suitable for pirated tracking when distributing videos on a large scale.