| In the21st century, with the development of multimedia, the technologies of videocommunication and network transmission have been widely used. As an intuitionaland abundant information carrier, the video has been exerting an increasinglysignificant impact on the human communication and the social development. First ofall, video communication is replacing the single voice and text communication step bystep in the modern society. The advanced techniques, like the3G mobile videocommunication, the remote video conference and the remote video surveillancesystem etc, have become the very useful communicating approaches and have acted asthe powerful “weapons” to ensure the social stabilization. Secondly, in terms of thepersonal entertainment, the explosive video information has dominated people’s dailylife and people’s need for information has transformed from the single text to thevideo, for example, Youku and YouTube are the well-operated network videoplatforms. While the post-PC era is arriving, the pursuit for the mobile terminals andthe demand of video applied at mobile terminals are increasingly large. Moreover,with the support of information technology development, consumers’ expectation forthe video experience is higher and higher. Therefore video communication is currentlyfacing the unprecedented chances and challenges.Video coding/decoding is one of the key techniques that ensures informationeffective transmission and storage and a significant part of the cutting-edge information technology. H.264/AVC (hereinafter referred as H.264) is the videocoding standard jointly made/recommended by ITU and ISO/IEC and the part10ofthe MPEG-4standard. In March2013, the ITU and ISO/IEC released the mostupdated standard i.e. HEVC (High Efficiency Video Coding). HEVC is designed forthe high resolution video format, based on the H.264framework and makes theimprovements on each part of encoding and decoding to meet the special demand ofhigh resolution. As a mainstream coding standard, H.264/AVC brings in a number ofadvanced techniques. Compared with the previous encode/decode standards, theaspects on coding efficiency, network adaption and compression quality, etc, havebeen obviously improved. However, the RDO (Rate Distortion Optimization) broughtby H.264/AVC causes the chicken-and-egg dilemma. Consequently, the rate controlof H.264becomes more complicated than the previous encode/decode standards. Thisdissertation proposes a novel rate control algorithm based on the HVS by a series ofexperiments and analyses on the standard rate control algorithm (JVT-G012).This dissertation starts with the description of the basic principles of H.264/AVC,such as intra-, inter prediction, DCT transform and quantization coding, together withthe overview of the basic concepts of H.264/AVC standard.Besides, it introduces a H264/AVC compressed domain video segmentationalgorithm which uses the macro-block motion information extracted from H.264bitstream to cluster the macro-blocks and partition the macro-blocks areas. Beforeusing the macro-block motion information, the reliability analysis and thetwo-dimension spatial filter processing are respectively carried on the original motionvector field and the value field of macro-block partition, so that a more reliablemotion field and macroblock weight field are obtained. Ultimately, mean shiftclustering algorithm is carried on the reliable motion field to achieve the movingobject extraction.In the constant channel environment and the current existing rate control schemecondition, in order to improve the quality of video coding, this dissertation conducts anumber of experiments, analyses and proposes a novel rate control scheme based on the JVT-G012. The proposed rate control scheme inherits the basic framework ofJVT-G012(GOP layer, Frame layer, Basic layer bit allocation and MAD predictionmodel etc.) and improves the MAD prediction model, the frame layer bit allocationand the Basic unit bit allocation.To obtain the more accurate value of MAD, this dissertation utilizes the spatialand temporal correlation adequately proposing an adaptive temporal-spatial MADprediction model. In JVT-G012rate control scheme, the frame bit allocation isachieved by setting a target buffer level. However, in proposed rate control scheme,the parameter is corrected and extended from single proportion controller toproportion-integral-differential controller (PID controller). Eventually, combined withthe proposed region segmentation algorithm, the basic unit layer bit allocation isadjusted. For the limited bandwidth application environment, the performance of theproposed rate control scheme is highly improved in MAD prediction and bitallocation. The experiment results show that the proposed rate control scheme forH.264/AVC increases the subjective quality of reconstructed video efficiently. |
PDF Full Text Request