Research On Robust Video Information Hiding Algorithm For Reducing H.264/AVC Distortion Drift

With the rapid development of the Internet age,people pay more attention to information security.Due to the popularity of H.264/AVC compression algorithms,video steganography based on H.264/AVC has attracted attention.In the process of steganography,not only will the distortion of the data be embedded,but also different types of passive attacks may occur in the transmission process,which will affect the implementation of the steganographic algorithm.This paper proposes a robust video steganography algorithm to reduce the distortion drift between H.264/AVC frames.By analyzing the reason why video steganography generates distortion drift,a solution to reduce distortion drift has been proposed.Perfonning entropy decoding on H.264/AVC to obtain the 4x4 DCT coefficient matrix of the motion vector residuals in inter-luminal macroblocks of P-frames:The coefficients in the matrix depend on the size of the distortion,according to the order of priority from high to low the the coefficients in the matrix is divided into three levels C₁,C₂,C₃.According to the embedding algorithm proposed in this paper,the privacy data is in turn embedded into the higher priority DCT coefficient set.The experimental results show that the proposed algorithm not only can effectively control the distortion drift to improve the video quality,but also further increase the embedded capacity.By learning the convolutional coding,the idea of convolutional coding is introduced on the basis of reducing distortion drift to improve the ability of video to resist passive attacks in the transmission process.The preprocessing of the convolutional encoding of the private data forms a new data stream to be hidden and is embedded in the carrier video.After the receiver extracts the privacy data,Viterbi decoding operation is performed on the extracted data to obtain the original embedded private data.Experimental results show that even when the convolutional code length is greater tharn 6,even if there is a bit error rate of 10%,the private data can still be completely extracted.Experiment reSsult shows:Differentiating the carrier coefficients with different priorities can reduce the distortion caused by data embedding in each frame;Processing each frame in reverse order can reduce the accumulation of distortion drift;Improve the embedded algorithm to further increase the embedded capacity;Adding convolutional coding to the embedded private data can improve the robustness of the video during transmission.