Study On Detection Of Prediction Mode Modulated Information Hiding For HEVC

With the development of the new generation coding technology-HEVC,more and more video files using HEVC as the encoding technology,and transmit on the Internet.Because of the excellent characteristics of HEVC technology,more and more HEVC vedios are uesd as the carrier of information embedding by the information hiding technology.With the development of information hiding technology,illegal information hiding operations also pose a serious threat to the security of the Internet Environment.Therefore,the research on the detection algorithm of HEVC video information hiding is of great significance for the maintenance of network security.FIrst of all,this paper introduces the background of video steganalysis Technology,then the research status at home and abroad are described in detail and the necessity of this study is pointedout.The following work has been done for the detection algorithm of HEVC video information hiding:Firstly,the principle of HEVC video codingis analyzed.This chapter introduces the main links of HEVC coding,such as intra-prediction,inter-prediction,transform and quantization,entropy coding and so on.Then,the part of intra-prediction is introduced in detail andanalyzes the new characteristics of intra-prediction coding in HEVC,the calculation of pixel value,the selction of intra prediction mode,which lays the foundation for the later analysis of the information hiding algorithm by modified intra prediction mode.Secondly,a steganalysis algorithm based on the correlation of prediction mode is proposed,which based on the analysis of the steganography algorithm by modify intra prediction mode.Above all,the influence of the information embedding on the video prediction mode is pointed out.Secondly,the correlation characteristics of theprediction mode are constructed,and the experimental results show that the information hiding has a great influence on the correlation of the prediction mode.Finally,the video features are input into the SVM classifier for experimental analysis.The experimental results show that the proposed method can achieve good detection performance by using the correlation feature of the intra prediction mode,and the detection rates of videos with different resolutions and different contentsis are higher than 96%.Thirdly,the selection of the size of prediction unit(PU)in the HEVC intra prediction is analyzed in detail,then an algorithm based on the change rate of the prediction unit before and after compression is proposed.First of all,the influence of the modified prediction mode on the partition of the prediction units size is analyzed.Then,the experimental results show that the prediction units changed after the information embedding objectively.By using the idea of calibration,the video with secret informationis compressed with the same parameters.It is found that there is a tendency for the prediction units to recover from the original video by the result.According to the characteristics of this phenomenon,the experiment was carried out.The experimental results show that the proposed algorithm has excellent detection performance for different resolutions and different contents of videos,and the accuracy rate can achieve more than 98% of high resolution video,moreover,the complexity of algorithmis low.