Detection Of Double Compression For HEVC Videos With The Same Coding Parameters

The rapid development of mobile communication systems has made digital video the most popular information carrier.However,due to the ease of operation of video editing software,the integrity and authenticity of digital video cannot be guaranteed.Double compression for digital video file is a necessary procedure for malicious video content modification,which means the detection for double compression is also an important auxiliary measure for video content forensics.The detection for double video compression can greatly reduce the cost of video forensics.High Efficiency Video Coding(HEVC)is the latest standard of video compression.The HEVC video is a video whose codec reduces the compression rate by half when compared to H.264/AVC(Advanced Video Coding).So it supports ultra-high resolution videos.Nowadays,mobile phones and surveillance cameras have supported HEVC videos.It is of great significance to detect double HEVC video compression.Double compression detection with different coding parameters has achieved a high accuracy,but in the compression process with the same coding parameters,the trace of HEVC video double compression is very small and the detection is much more difficult.This paper proposes two detection algorithms for video double compression with the same coding parameters from the process of prediction and transformation respectively.Both algorithms are feasible and have high detection accuracy.The video double compression detection algorithm based on PU is proposed to expose HEVC video re-compression.The quality of video decreases and tends to unchanged under the same encoding compression for many times.In this process,the change of PU mode of I frame and P frame in the same position reveals the law of quality degradation.The degree and trend of quality degradation can distinguish between single compressed video and double compressed video.The average different PU modes of the n th compression and the(n+1)th compression of each I frame and P frame at the same position are counted to form a 6-dimensional feature set,which is sent to a support vector machine(SVM)for classification.The average accuracy of the algorithm detection is over 95% and it can reach a higher rate by different video resolutions,GOP settings and frame deletion.In order to improve detection accuracy,the video double compression detection algorithm based on TU is proposed to expose HEVC video re-compression.Selecting TU as the research object can reduce the fluctuation caused by the change of PU prediction mode in the prediction process.The TU mode is easier to converge than the PU mode.The average number of changes in the n th compression and the(n+1)th compression at the same position for each I frame of the four different sizes of TUs is counted to form a 12-dimensional feature set,which is sent to a SVM for classification.The algorithm improves the detection accuracy and has a strong generalization ability.