The Research Of Video Predictive Code And Error Resilience Based On H.264

H.264video coding standard is proposed by ITU-T and ISO in Mar2003, the standardgot much attention with its favorable network friendliness and high coding efficiencycompared with the previous video coding standards. The original design target of H.264/AVCis its coding efficiency double than H.263and MPEG-4with the same decoded quality, inorder to better realize to transmit video stream under low bit rate environment.In order to better implement the transmission of video stream under low bit rateenvironment, H.264/AVC adopts a series of new technology. Although the adoption ofadvanced technology obviously improves the performance of coding and decoding, butincrease the complexity of module and also exponentially increase the complexity ofalgorithm, which limits the application of H.264/AVC in the real-time video communicationand in a way, and inter-frame search algorithm is the most complex and mosttime-consuming part in the process of video encoding, so this paper mainly focuses onresearching video frame search algorithm and simple error resilience.The first part of this paper devotes the H.264framework and its key technologies, andemphatically elaborates the inter-frame predictive encoding, which detailed illustrate thecomplexity of H.264, thus confirm the necessity of researching inter-frame fast searchalgorithm.The second part of this paper emphatically elaborates the H.264inter-frame fast searchalgorithm. Firstly based on the diamond search algorithm, improving it, through thesimulation, the paper gets expected result. Then using the improving diamond searchalgorithm as premise, the article proposes BMV-AES algorithm. The algorithm which basedon motion vector interdependency, for macro-blocks of video image which meet certaincondition could terminate search algorithm in advance, and predict motion vector by relativemacro-blocks; for macro-blocks which fail to terminate search algorithm will adopt differentmacro-block search strategies according to different motion type, as for small motion typeadopting8*8macro-block size with cross search strategy, and as for large motion typeadopting4*4macro-block size with improved diamond search strategy. This algorithmcompare with traditional single mode improving diamond search algorithm and another spatio-temporal interrelated motion vector field adaptive search algorithm in the article, thepaper raises a conclusion that the performance of BMV-AES algorithm is excellent throughthe simulation. Finally, according to the properities of H.264error resilience, separatelyreaching error resilience by the estiomation of lost macro-blocks motion vector on MVFASTand BMV-AES algorithms, we conclude that the error resilience of BMV-AES is a little betterthan MVFAST.