Detection Of Information Hidin In H.264/AVC

H.264/AVC is widely used in various fields for its outstanding qualities, while that also makes H.264/AVC video stream a good carrier for information hiding. But illegal and malicious information hiding will make an extremely threatened to the information security. Therefore, it is significant to study the steganalysis algorithms of the H.264/AVC information hiding methods for the authentication, encryption and national information security in the information transmission. In this paper, the main contributions to the steganalysis algorithms of H.264/AVC information hiding can be summarized as follows:1. The algorithms of H.264/AVC information hiding have been investigatedAccording to some special encoding characteristics in H.264/AVC, we summarized five types of information hiding methods. The next step is based on it. For the information hiding methods based on modulating intra prediction modes and modulating the coefficients of integer DCT transform, we have studied in detail and gave the theory basis of them.2. A steganalysis method of information hiding based on modulating intra prediction modesFirstly, we analyzed the methods of information hiding based on modulating intra prediction modes, and included that the embedding methods and extraction methods of them are similar. Secondly, we analyzed the changes of remarkable characteristics in intra-frame coding caused by modulating intra prediction modes for information hiding. We found that the inherent correlation among the prediction modes in different luminance blocks, belonging to an intra-frame coding macroblock, was changed. According to several different positional relationships of the adjacent blocks in spatial domain, we have designed statistical models corresponding to the correlation of prediction modes to make quantitative extraction of these correlation characteristics. And a steganography detector was constructed based on the support vector machine. Based on the steganography detector constructed in this paper, the experimental results show that, for different test video sequences in five different feature vectors, the mean of the detection accuracy and precision ratio is more than89.69%and86.25%respectively when the embedding rate is larger than12.5%, and the recall ratio is also more than96.42%.3. A steganalysis method of information hiding based on modulating the coefficients of integer DCT transformFirstly, we analyzed the methods of information hiding based on modulating the coefficients of integer DCT transform. Secondly, we analyzed the significant feature changed before and after the information hiding. We found that the coexistence characteristic of the residual matix’s coefficients was changed. And then combined with the support vector machine, we have implement this type of information hiding method. The experimental results show that, for different test video sequences, the mean of the detection accuracy and precision ratio is more than93.33%and96.12%respectively when the embedding rate is larger than12.5%, and the recall ratio is also more than88.33%.Based on investigating the algorithms of H.264/AVC information hiding in detail, we have then designed the corresponding steganalysis methods of information hiding based on modulating intra prediction modes and modulating the coefficients of integer DCT transform.