3DTV Video Error Concealment Technology Research For Transmission

With the rapid development of wireless network and multimedia technology, three-dimensional (3D)video has been gradually entering the leading ranks of the multimedia video. However, packet loss and bitdistortion occur constantly in network transmission channel as a consequence of the great amount ofinformation encoded and the limits of decoding time and network bandwidth. They may invoke displayingerroneously because of decoding normally. Thus, it is particularly significant to study the three-dimensionalvideo error concealment technology. The error concealment technology mainly recovers the lostinformation using the neighboring correctly received information at the decoder side. Traditional errorconcealment technology utilizes time domain correlation and viewpoints domain correlation of videosequences to restore the lost information, but it’s still needed to further ameliorate recovery effect.Therefore, this paper puts forward three algorithms as follows:(1)To address that existing error concealment methods fail to fully consider the human subjectiveperception for full frame loss. This paper presents an error concealment algorithm based on structuralsimilarity (SSIM) which reflects the human eye’s subjective perception for the full frame loss. Firstly, ajudgment approach for different macroblocks’ reference models based on SSIM is put forward, and themacroblocks are divided into time domain reference and viewpoints domain reference. Secondly, thepreviously derived models before the current lost frame are taken as the corresponding macroblock modesin the lost frame. Finally, the different lost macroblocks are restored using motion compensation prediction(MCP) and disparity compensation prediction (DCP), respectively. Experimental results show that theproposed algorithm of average peak signal to noise ratio (PSNR) can improve2.26dB compared with thatof traditional error concealment method, and subjective quality has yielded corresponding improvement.(2)To address that B frame macroblock loss in stereoscopic video with hierarchical B pictures (HBP),a novel error concealment algorithm is proposed. Firstly, in order to reduce complexity, the lostmacroblocks are classified into rapid motion and slow motion types according to the image contentinformation. Secondly, the structure of the encoding HBP is analyzed and discussed, and error concealmentmethod based on the time-domain correlation and error concealment method based on block matching areproposed to restore the two different types of macroblocks. Experimental results show, compared with thatof traditional error concealment algorithm, the proposed algorithm significantly improves the objectivequality by an enhancement of PSNR from0.23dB to1.26dB and their subjective quality.(3)To address that stereoscopic video decoding display in the terminal equipment, the paper designsthe stereoscopic video player (SVP) and analyzes the application of SVP. Firstly, the design progress andfunction feature of the SVP are analyzed based on the H.264video codec standard and stereoscopic videodisplay principle. Secondly, the SVP is applied to error concealment technology and quality evaluation, theperformances of different error concealment methods are compared from the perspective of decoding timeand the subjective quality evaluation. Experimental results show that SVP earns better display effect andlower cost, so the SVP tends to have a broad application prospect.