Research On Video Steganography Technology Against H.264 Compression

With the popularization of mobile smart terminals and the reduction of network tariffs,the proportion of video traffic in the network is increasing day by day,providing a rich cover for video steganography.In order to save storage and transmission costs and enhance user experience,many platforms re-compress and encode videos.Among the many video compression coding standards,the H.264 standard stands out by virtue of its advantages such as low bit rate,high quality,strong fault tolerance and strong network adaptability,and has become the most commonly used standard for video compression.However,after many hidden videos generated by existing video steganography algorithms are compressed by H.264,the secret information cannot be extracted correctly.Therefore,studying video steganography against H.264 compression has important practical significance for enriching steganography methods and improving the security of covert communication.Based on the systematic introduction of the research status,the thesis carried out research on two existing problems,and proposed two video steganography algorithms against H.264 compression.The specific work is as follows:1.Aiming at the problem that the existing video steganography with motion vector as the embedded domain cannot resist H.264 compression,an anti-H.264 compression video steganography algorithm based on block value replacement is proposed.First,the pre-compressed carrier video obtains the original motion vector in the embeddable area,and embeds the encoded secret information into the motion vector by replacing the least significant bit;predicting the brightness value of the coded block with the corresponding secret motion vector The brightness value of the block is replaced to minimize the absolute error sum between them;finally,the generated hidden video is compressed by H.264 to obtain the hidden H.264 video.Experimental results show that compared with the existing algorithms based on motion vectors,the algorithm has a lower bit error rate for extracting secret information and better visual invisibility.2.Aiming at the problem that the existing transform domain video robust steganography algorithm can modify the video frame greatly,an anti-H.264 compression video steganography algorithm based on the relative relationship of DWT coefficients is proposed.First,analyze the changes in the low-frequency DWT coefficient difference between adjacent frames of the video after H.264 compression;then,use the Gaussian function to model the change in the coefficient difference,and select the appropriate coefficient difference threshold according to the predefined extraction error rate;Finally,according to the selected threshold,the low-frequency DWT coefficient values of adjacent frames are modified in pairs to modulate the difference between them to realize information embedding.Experimental results show that the algorithm has high visual invisibility,and has better robustness against H.264 compression for videos that are not violent in motion.Finally,I summarize the work of the full text,point out its shortcomings,and make future research prospects for video steganography against H.264 compression.