Video Steganalysis Algorithm Based On Spatial-Temporal Local Binary Pattern

In recent years, with the rapid development of multimedia technology and computer network, digital products have brought much convenience to people's daily work and life, while at the same time also caused a variety of information security. As a research hotspot of information security field, information hiding has opened a new way for the storage and transmission security of multimedia information. Information hiding refers to embedding secret information into video without causing the suspicion of the third party, so as to realize the purpose of copyright protection and covert communication. As its against technology, video steganalysis refers to the detection process of identifying whether suspected carrier contains confidential information in the case of known or unknown embedding algorithm, which starts late and is very significant for the development of information security technology.Spread spectrum embedding is a typical video steganography in information hiding field with the advantages of resisting video compression. To complete the detection of spread spectrum information effectively, a video steganalysis scheme based on spatial-temporal local binary patterns(Spatial-Temporal Local Binary Pattern, ST_LBP) is proposed. considering the spatial-temporal correlation of video sequences, this method establishs three-dimensional orthogonal planes, extracts S_ LBP ?_xT LBP ?_yT LBP features from XY, XT, YT planes respectively and then does histogram statistics of these three planes. In order to weaken the influence of local motion and improve the detection performance furtherly, it also proposes the concept of activity factor on the basis of the block motion analysis. Finally, it selects original video sequence and does experimental simulation of ST_LBP steganalysis algorithm with Matlab to verify the feasibility and robustness of this algorithm. Experimental results show that, compared with the traditional Rich Model method, this model not only reduces the influence of local motion on detection precision, but also acquires a higher detection rate.