Research On Large Capacity H.264/AVC Video Information Hiding Algorithm In QDCT Domain

From the appearance of H.261 to the wide application of H.264,the iterative update of digital multimedia technology brings not only convenience to people,but also the problem of cyber space security.In order to solve the problem of copyright disputes and information leakage in cyberspace security,combining the information hiding technology and video compression coding is an effective mothod to solve this problem.Thus,the research on H.264 video information hiding has caused wide public concern.In this thesis,we focus on studying data hiding technologies in H.264/AVC video.The main contributions are listed as follows:1.A data hiding algorithm for H.264/AVC video streams with low bit-rate growth based on three-dimensional(3D)space is proposed.The open-loop data hiding in quantized discrete cosine transform(QDCT)domain can easily cause drift distortion,and the intra-frame distortion drift is an important factor in sharply decreased stego-video quality.Therefore,under the condition of averting the distortion drift,the algorithm embeds B-ary secret data by encoding 4×4 blocks' coefficients to be embedded in the three-dimensional space.In the extraction process,B-ary secret data,which use the intra-prediction model of the currentembedded block and adjacent blocks with the QDCT coefficients,can be correctly extracted by computing the 3D code.Experimental results have shown the proposed method can reduce the bit-rate growth of the stego-video streams(for a fixed payload)as well as improve the capacity of the hidden information,when compared with some related works.2.An information hiding scheme based on multi-classification of reference pixels in QDCT domain for H.264 video is proposed.Under the condition of averting the distortion drift,the reference pixels of the adjacent blocks are first classified into multi-groups.Then,9-ary exploiting modification direction(EMD)embedding algorithm and matrix coding are used to obtain a high embedding capacity by fully exploiting the available QDCT coefficient-pairs of the 4x 4 blocks of I frames.Meanwhile,the quality of stego-video can be ensured by changing,at most,one set of QDCT coefficients.Furthermore,the detection of hidden information only requires to decode the intra QDCT coefficients and intra-prediction mode without decoding the video completely.The secret information can be extracted completely.Experimental results have shown that the proposed method can greatly improve the capacity of hidden information,and have little impact on the bit rate and quality of video.