An efficient stereo video coding technique based on H.264

• Main Author: Kiran Nanjunda Iyer.
• Other Authors: Tan Yap Peng
• Format: Theses and Dissertations
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Computer science and engineering::Data::Coding and information theory
• Online Access: http://hdl.handle.net/10356/18837
• Institution: Nanyang Technological University

Compression of stereoscopic and multiview video data is important because the bandwidth necessary for storage and transmission linearly increases with the number of camera channels. Monocular compression schemes such as MPEG-2, MPEG-4, and H.264 achieve compression by decorrelating in the spatial and/or temporal correlation domains. Stereo video streams add a new dimension, namely the correlation between the left and right images. Direct application of any monocular compression schemes is suboptimal because the decorrelation is limited to spatial and temporal correlation domains. An efficient stereoscopic coding scheme is presented in this document. This method achieves compression by extending the decorrelation process of monocular compression schemes to spatial, temporal and interview correlation domains. The proposed scheme also supports asymmetric view coding which would exploit the psycho visual redundancy and improve the compression efficiency. The proposed method requires only tools corresponding to the baseline profile of H.264 and hence operates at a very low complexity. It is relatively simple to implement the stereo codec both in hardware and software. Experimental results prove that the proposed stereoscopic coding scheme achieves a higher PSNR (objective image quality) at lower bit rates as compared to other low complexity stereo coding schemes such as the simulcast and transposed picture ordering. The results of asymmetric view coding using the proposed method indicate the improvement in the perceptual quality of the image at lower bitrates as compared to conventional asymmetric coding schemes.