Research Of Video Encoding Schemes In Energy Limited Environment

With the rapid development of wireless networking and mobile systems, video streaming in wireless environment has attracted more and more focus and research. H.264, the newest video standard, is suitable for wireless environment, because it has very high compressing quality and can be robust in network transmission. However, it has a rather high computational complexity in the encoding and decoding process. It is difficult to perform real time application with the reference software provided by ITU. So we have to optimize the process in the encoder and decoder to decrease the complexity.So far the video encoding technologies are mostly considered to reach higher compressing quality, so there is much space to reduce the complexity for H.264 encoder. By analyzing the encoding process, we can find that the inter prediction is the most consuming module, and the variable block size motion estimation adopted by H.264 intensify the complexity greatly in inter prediction. After researching and analyzing the methods, two new schemes are proposed.Firstly, the low complexity mode decision algorithm based on motion region partition was proposed to reduce the complexity of H.264 encoder by using the characteristics of spatial correlation. The proposed algorithm is to simplify the mode decision of inside macroblocks surrounded by identical motion vectors or high speed motion vectors. Experimental results show that the proposed algorithm reduces much computational complexity compared to conventional mode decision, but influences the compressing performance obviously in the meanwhile.Secondly,the fast variable block size motion estimation algorithm Using Geometric Structure Categories is proposed to reduce the complexity for the encoder by using the information of the current encoding macroblocks. The proposed method firstly makes mode decision and then estimates motion vectors to effectively reduce the complexity of motion estimation. Experimental results show that the proposed algorithm reduces much more encoding time, while maintaining the rate-distortion performance comparable to that of exhaustive conventional method.So many existing software and hardware techniques have been proposed to design a low-complexity encoder, but when we consider the applications in the wireless environment, another problem appears. Portable video communication devices, powered by batteries, have limit energy supply. Therefore, one of the central challenging issues in portable video communication system design is to minimize the energy consumption of video encoding so as to maximize the operational lifetime of devices. In this paper, we propose a general method to save the data processing energy and further reduce the energy consumption significantly using power-rate-distortion optimization.