Complexsity Quantization And Adptive Control Of HEVC Decoding

In recent years, with the rapid development and maturity of digital video applications industry chain, HD,3D, mobile wireless has become a main trend of the future development of video applications. This means that the future development of video technology will focus on the development of the videos with high definition, high frame rate and high compression ratio. Based on that, a new generation of video coding standards HEVC which means highly efficient video coding standard came into being.As a new generation of video coding standards, HEVC not only inherits the advantages of the previous standards, but also integrated with a variety of state-of-the-art video coding technologies, and it is designed to achieve the ultimate goal of doubling the compression efficiency of H.264/AVC high profile.However, the significantly increase of HEVC compression efficiency sacrifices the complexity of HEVC encoding and decoding algorithms, at the same time as HEVC often process the videos with high definition and3D content, compared with the H.264the complexity of HEVC encoding and decoding algorithms increases greatly, which undoubtedly makes the processing capability and the optimization consumption capability of limited resources of video decoding terminals, especially of mobile wireless terminals confront with greater challenge.Based on that, this paper puts forward a quantization model on HEVC decoding complexity, and it estabilishes HEVC decoding complexity model by choosing complexity quantization parameters in intra-prediction decoding process and inter-prediction decoding process. In other words, it estimates overall HEVC decoding complexity by quatizing the complexity of intra-prediction decoding process and inter-predictiopn decoding process.Besides, based on the quantization of HEVC decoding complexity we put forward with an adaptive control algorithm on HEVC decoding complexity in this paper. This algorithm does not need to make any changes to the processors of the existing video terminals, but only need HEVC encoders generate different sequences according to the processing capabilities and the amount of available resources of different terminals, in the mean while,HEVC encoders must ensure the minimal RD-cost of the generated sequences. In other words, the algorithm embedded decoding complexity quantization and adaptive control module in HEVC encoder side, which aims to satisfy the processing capabilities and the need of resources saving (such as reduce power consumption) of different terminals, and to ensure the quality of decoded videos at the same time.At the same time, this paper gets rid of the quantization mechanism by using time in previous complexity quantization methods, but use cache access count of CPU during HEVC decoding to overcome the shortcomings when using time as complexity quantization parameter.Expreiments over divierse videos the proposed complexity quantization and adaptive control algorithm of HEVC decoding can reduce about35%of the decoding complexity for the sequences, and ensure the PSNR only reduce less than1dB at the same time. Based on that the HEVC encoder side equipped with complexity quantization and adaptive control module can effectively balance the bit rate distortion and decoding complexity of the sequences at the same time.