Researches On Rate Control Algorithm For Stereo Video Coding

As the rapid development of network, video communication technology and the increasinglygrowth of human visual needs, stereoscopic video technology is going into human’s life. Ratecontrol is one of key technologies in stereoscopic video coding and transmission. The study themeof this thesis is researches on rate control technologies for stereoscopic video coding.Firstly, the principle of imaging and display for stereoscopic video was introduced, then, thetheoretical basis for stereoscopic video coding was stated, which includs the necessity ofcompressing stereoscopic video, basic methods of stereoscopic video coding and H.264-basedstereoscopic video coding algorithm.Secondly, based on the study of single-channel video rate control technologies, a linear modelbased I-frame rate control algorithm was proposed, against the problem that there is not effectiverate control for I-frame in traditional rate control. By taking scene change, image complexity andcurrent buffer fullness ratio into account, quantization parameter for I-frame was calculated.Experimental results show that, compared with JVT-G012rate control algorithm, the proposedalgorithm had greatly improved the peak signal to noise ratio and rate-distortion performance whilethe output bit rate almost unchanged.In addition, according to characteristics of stereoscopic video, by taking image complexity,position significance and reference property between the left and right pictures into account, a newH.264-based rate control algorithm for stereo video coding was presented. Experimental resultsshow that the proposed method can efficiently implement rate control for stereoscopic video.Compared with the quantization parameter (QP)–fixed mode method, the proposed method hadgreatly improved the peak signal to noise ratio and rate-distortion performance, and the averagePSNR gain between rate-distortion curves (BDPSNR) is0.21dB.Finallly, based on summarizing the study status of rate control for multi-view video coding(MVC), an initial QP-estimated based rate control algorithm for MVC was proposed according tothe hierarchical structure of MVC. Experimental results show that the proposed algorithm canefficiently implement rate control for multi-view video. Although the proposed algorithm isinsufficient from the peak signal to noise ratio perspective, it provides a good idea for the researchin this field.