The Complexity Control Technique For H.264 Encoder Based On Wireless Mobile Terminals

With the fast development of computer techniques and information processing techniques, the newly video coding standard—H.264 was proposed by the moving picture expert group (MPEG) and ITU telecommunication standardization sector (ITU-T) in the year 2003. H.264 has adopted a number of advanced coding techniques to improve the coding efficiency, and also has a friendly network applicability, which can meet the requirements of various video applications. Therefore, H.264 can be widely used in video communication, streaming media and digital broadcast and so on.However, the enhancement in coding efficiency of H.264 has been achieved at the cost of a considerable increase in computational complexity. A wireless mobile terminal (WMT) can not afford such a high complex video coding algorithm to operate for a long time because its battery is energy limited. Therefore, how to figure out the conflict between the huge power consumption of encoder algorithm and the limited battery energy of the WMT, so as to find an optimal trade-off between the coding performance of the encoder and the limited power/complexity of the WMT, has become a hot topic in wireless multimedia communication fields.In this paper, the research on H.264 based complexity control technique for wireless video transmission is carried out on WMT. The main researches include:First, we analyze the complexity of the H.264 encoder deeply, and defined the measurements of the intra and inter prediction coding complexity.Then, an efficient complexity allocation and control (CAC) algorithm for intra prediction coding in H.264 is proposed, which firstly rearranges the candidate intra prediction modes with priorities based on the analysis of a macroblock (MB) edge-direction, so as to allocate the available complexity to the modes with high priorities first when perform the complexity control. This can achieve an efficient complexity allocation and control for intra prediction coding with minimum degradation in video quality.Next, an efficient complexity allocation and control algorithm for H.264 inter prediction coding is proposed. Based on the texture and edge-direction analysis of an MB, this algorithm firstly rearranges the candidate inter prediction modes with priorities; Then, a newly rate-distortion-complexity (R-D-C) optimization cost function is proposed by combining the complexity parameter with the traditional rate-distortion (R-D) cost function, which can make a trade-off between the coding performance and complexity. By using this algorithm, the encoder can adaptively allocate the available complexity to each inter prediction mode according to its priority, and determine whether it is valuable performing the sub-block partitions and sub-pixel motion vector search, so as to control the complexity of inter prediction coding efficiently.Afterward, according to the energy character of the WMT, the complexity scalable control (CSC) algorithm for H.264 encoder is produced by combining the proposed CAC algorithms together, which can adjust the H.264 coding complexity at scalable levels adaptively according to current energy condition of a WMT. Experimental results show that our algorithm can maintain the R-D performance of the encoder when the coding complexity is controlled at low levels, so as to prolong the operational lifetime of the battery for hand-held devices.Finally, we propose a video transmission system based on the PDA (Personal Digital Assistant) platforms over wireless local area network (WLAN) to validate and perform the proposed algorithm.