Research On Theories And Methods Of Digital Video Information Hiding

Information hiding technique has a long history.In the digital era,information hiding is a technique which embeds data into digital media such as texts,audio,images and videos.The embedded data can be extracted from the stego media when the data embedding algorithm and the data embedding key are known to users.As a practical technique,information hiding plays a significant role in maintaining the digital media content security.Currently,videos which have become one of the most prevalent media in networks are accessible covers.However,compared with image information hiding,video information hiding still remains in the preliminary research stage and many key problems deserve investigation.Therefore,it is meaningful and valuable to study ef-fective theories and methods of digital video information hiding.Based on the analysis of the video information hiding theories,this dissertation investigates corresponding video information hiding methods.The main work and in-novation points of this dissertation are listed as follows.1.Propose the reversible information hiding method for adaptive video error con-cealment.To enhance the video transmission robustness in error-prone networks which suffer bit errors,packet losses and malicious attacks,the reversible information hiding method for adaptive video error concealment is proposed.The Laplacian probability distribu-tion of DCT coefficients in videos is analyzed and verified.This theory demonstrates that the sufficient embedding capability can be obtained if messages are embedded in quantized DCT coefficients of videos.Furthermore,this theory can be utilized in the following research on video information hiding.In this proposed method,macroblocks whose motion intensities are larger than the predefined threshold are allocated to the region of interest and motion vectors of macroblocks belonging to the region of interest are embedded into the background region within the same frame.The histogram shift-ing technique based on quantized DCT coefficients of coded macroblocks is used to achieve the lossless recovery of the background region after data extraction.If a mac-roblock in the region of interest is found missing at the decoder,the embedded motion vector can be extracted from the corresponding stego macroblock for the recovery of the missing macroblock.Experimental results show that the better error concealment performance can be achieved compared with existing related methods.2.Propose the video steganography method based on the minimal embedding dis-tortion principle.According to the minimal embedding distortion principle,the effective method for motion vector-based video steganography is proposed.To resist the detection of ste-ganalysis,this method aims to design a suitable distortion function expressing the em-bedding impact on motion vectors by exploiting the spatial-temporal correlation.The practical embedding algorithm is implemented using syndrome-trellis codes(STCs).T-wo factors which are the statistical distribution change(SDC)and the prediction error change(PEC)are considered in the proposed distortion function.Based on the min-imal embedding distortion principle,the motion vector reversion theory during video recompression is analyzed.This theory demonstrates that it is necessary to consider the prediction error change(PEC)for designing the distortion function.Experimen-tal results show that the proposed method can enhance the undetectable performance significantly compared with existing related motion vector-based video steganographic approaches and can obtain the better video coding quality.3.Propose the reversible information hiding method of decreasing inter-frame distortion drift in encrypted videos.Due to the privacy-preserving requirement for cloud data management,it is nec-essary to store encrypted videos in the cloud server.To address the authentication of encrypted videos in the cloud server,the reversible information hiding method of de-creasing inter-frame distortion drift in encrypted videos is proposed.In the encryption phase,the codewords of intra prediction modes,motion vector differences and partial residual coefficients are encrypted without video bit rate increment for preserving the confidentiality of video content.In the data hiding phase,the theoretical analysis of the picture distortion caused by data embedding and the subsequent inter-frame distor-tion drift is firstly presented.Based on the analysis,the embedding distortions caused by modifying different residual coefficients are estimated and data are embedded into residual coefficients with different priorities for decreasing the inter-frame distortion drift.The data embedding is implemented by the bidirectional histogram shifting tech-nique.Experimental results verify that the proposed method can obtain better decoded reconstructed video quality compared with existing related methods.4.Propose the reversible information hiding method for texture videos and depth maps coding.To support 3-D video and free-viewpoint video applications,efficient texture videos and depth maps coding should be addressed.The reversible information hiding method for texture videos and depth maps coding is proposed for securing depth maps.This method utilizes the theoretical analysis of inter-frame distortion drift for embedding the depth video bitstream in corresponding texture video.At the sender end,the depth video bitstream obtained by depth down-sampling and compression is embedded in residual coefficients of corresponding texture video.At the receiver end,the depth maps can be retrieved with scalable quality after data extraction,video decoding and texture-based depth reconstruction.Due to the attractive property of reversible data hiding,the tex-ture video bitstream can be perfectly recovered.Experimental results demonstrate that the proposed method can achieve better video rendering quality and coding efficiency compared with existing related schemes.