Research On Scalable Video Coding Technology Based On HEVC Standard

With the increasingly popularity of video information and the emerging variousvideo applications, video coding technologies have achieved rapid developments inrecent years. Due to the inherently heterogeneous networks of Internet, differentdevices show various capabilities on processing and displaying videos for userterminals, which also greatly challenges the transmission and applications of multi-media contents. Scalable Video Coding technology is widely investigated due to itsability to handle that problem.By the corporation of the VCEG organization in the ITU-T and the MPEGorganization in the ISO/IEC, the newly designed video coding standard, high efficiencyvideo coding standard, is successfully published at Jan.2013. The HEVC standard stilladopt the compound coding framework and several advanced coding tools areincorporated into the HEVC standard. It aims at saving about50%bit-rates comparedwith the precede H.264/AVC standard under the same visual quality. This paper usesthe HEVC standard as an anchor and designs the scalable high-efficiency video codingunder the HEVC framework. We study the features of the SHVC standard and proposeseveral schemes to improve the enhancement layer under the corresponding testsoftware.The base layer and the enhancement layer of quality scalability have the samevideo resolution and frame rate. In order to achieve the visual quality scalability,different quantization parameters are needed for different layers. The base layer is stillcoding according to the HEVC standard. For the enhancement layer, make full use ofbase layer information to guide the enhancement layer coding is the most outstandingfeature, we called inter-layer prediction. The inter-layer prediction mechanism canmake the enhancement layer prediction more accurate, so that the coding efficiency isimproved significantly. In this paper, a deep study on the mechanism of inter-layerprediction, we proposed several inter-layer prediction schemes, including inter-layertexture prediction, inter-layer motion prediction and so on. Without increasingthe coding complexity largely, we can achieve a high coding efficiency.The spatial scalability respectively uses the videos with different resolutions as theinputs of the base layer and the enhancement layer. Similar to quality scalability, spatialscalability also adopt inter-layer prediction mechanism to improve the coding efficiency.The difference is that we need apply up-sample processing on base layer reconstructed frame. Since up-sample is a lossy processing, thus inter-layer prediction mechanism ofthe spatial scalability is more abundant. We often requires auxiliary inter-layer filteringtechniques to reduce this loss of information. In this paper, we focused on the inter-layer prediction specificity of spatial scalability and then we made a considerable scaleexperiments and optimization algorithms. At last, several inter-layer predictionschemes are proposed, including adaptive combined weighted prediction, adaptiveinter-layer filter and so on. Results show coding efficiency of the enhancement hasachieved great improvement.Besides, a summarized description of HEVC and SHVC will be showed in thispaper, which is the start point of all the research works.