Research On H.264 Video Optimization Over Wireless Channel

Along with the rapid development of wireless communication and multimedia technology in the world, video communication over wireless channel is becoming a hot issue in the research area of image communication. However, it brings up a huge challenge than others due to the inherent characteristic of the wireless channel. This dissertation faces the new developed video coding standard H.264/AVC, and operates research surrounding the wireless video compression and transmission. In order to provide high quality video, many technical issues are taken into consideration including how to reduce the computational complexity of encoding, how to transmit video data over wireless channel more reliably and robustly, rate control, and so on. The contents of our work are mainly divided as follows.First, according to the spatio-temporal correlationship of the neighbouring frame in sequences, a fast mode decision jointly with motion estimation is proposed. Mode decision is operated followed by block match searching only if the candidate block type is choosen as the optimized one. The experimental results show that the proposed algorithm reduces the encoding time without degrading encoding quality distinctly, equivantly increasing the negligible bit rate. On the other side, a dynamic threshold is applied in the judgment for improving the accuracy and robustness.Second, error resilient, as a key technology for reliable data transmission, is explored due to the bandwidth varing and high error ratio under wireless channel. In this dissertation, the error resilient is separated into tow part as follows. A scheme of data partition based on channel condition is proposed in encoder ender. The scheme chooses the work mode of NAL by channel feedback information. Then, the unequal error protection using RCPC is applied for the data partitioned data. Simulation shows that the trade-off between the coding efficiency and error robustness is obtained without additional bandwith and with high quality. On the other side, adaptive error concealment is proposed in decoder ender. The algorithm judges the motion degree for choosing the concealment by computing the number of the intra MB and acquiring the MV information.In the dissertation, the improvement rate control algorithm is given as follows. Bit allocation is improved by exploiting the spatio-temporal correlation sufficiently well for predicting MAD to estimate the unit complexity so that the accurary of rate control is achieved. Meanwhile, the algorithm employed HRD constrains for guaranteeing the buffer stationary. Simulations showed that this scheme improves the perceptual quality of the pictures with similar or smaller PSNR deviations.