H.264 Standard Video Decoder Optimization

With the development of information technology in 21st century, there grows the need of storing and transmitting video information efficiently. The property of digital video limits the expansion of multimedia services, so it is necessary to develop techniques for compressing these video data. H. 264 standard is jointly being made by ITU-T Video Coding Experts Group and ISO/IEC MPEG Video Group, named as joint video team (JVT). H. 264 greatly'improves coding efficiency and will be a new international video coding standard.The characteristic of H. 264 standard can be concluded mainly into three aspects. It is more practicable, more adaptive to IP net and mobile system and it improves the key components of coding structure based on hybrid video coding framework. For example, it adds coding mode decision, multi frame prediction, context-based adaptive variable length coding (CAVLC), 4x4 integer transform, loopfilter, etc. The improvement in performance leads to higher computational complexity in the new standard. The complexity of H. 264 encoder is about 3 times of H. 263 encoder and the complexity of its decoder is 2 times of H. 263 decoder. The optimization of H. 264 codec is important for its application in different fields.In this thesis, the optimization methods for video processing are analyzed with examples. Then, the omplexity of key modules in H. 264 decoder is analyzed and discussed. Optimization methods in integer transform, motion compensation and CAVLC is proposed and implemented based on general processor.The implementation of H. 264 decoder on handheld platform is also analyzed in the thesis. Some suggestion in optimization method based on certain processor is proposed.