Research On The Optimazation Of Encoding Algorithm Of The H.264/AVC Scalable Extension

Different compression and transmission systems have to be constructed for users in different conditios if they are built on the top of traditional non-scalable video encoding and decoding techiques.With the development of ubiquitous multimedia, scalable video coding techniques are desired for efficiently repurposing and transporting the video contects over heterogeneous network/devices. The acalable extension of H.264 is the newest work of developing joint standards for ISO/IEC and ITU. Based on the newest encoding and decoding techniques, temporal, spatial and quality scalability is realized with a unified framework in H.264 scalable extension standard at the cost of significantly increasing comlexity. In order to make full use of the emerging state-of-art technology, it is necessary to optimize it with deep reasearch. In this thesis, the focus is put on the study of coding structure, design of downsampling filter and the optimization of the streaming extracion.The thesis first reviews the standard of H.264/AVC analyzing the key mechanisms of encoding and decoding. Then, the thesis analyzes the framework, hierachical structure, encoding mode, application scenario and performance with the detailed experimental data. This thesis then focuses on scalable coding which can be classified into three types .With graded-depth study of key technologies, appropriate coding tools are selected to optimize coding algorithm. In the time domain, using the B layer Hierarchy to improve coding efficiency; in spatial domine, the design of spatial layer resampling process is focused on.Finally, in quailty scalabilty, the drift control problem is studied specially, simulation experiments indicate the use of a key picture based on the joint control structure of basic layer and enhancement layers will be able to effectively limit drift errors in a time-domain picture group and to ensure the efficiency of video encoding.The thesis then makes effort to optimize the extraction process; a quality-based rate-distortion optimization of hierarchical bitstream extraction method is given, with the comparison of the basic method, this approach can effictively tradeoff between image distortion and bit-rate.The thesis finally conducts a preliminary exploration to the combination of SVC with the mobile radio network. With the general analysis on the character of the3rd mobile radio channel and the transportation, the relationship of SVC and the enhanced network technique and protocols is described. The results show the benefit from using SVC and the new potentials for network operators and end user.