Research On Fast Algorithm Of High Efficiency Video Coding Based On HEVC Standard

With the rapid development of electronic display technology and mobile multimedia technology,high-definition video attracts the attention of more and more people since its abundant and sharp display ability,and people experience the visual immersion anywhere became a strong urge.In order to improve the efficiency of video compression,ITU-T and ISO/IEC established JCT-VC to develop a new generation High Efficiency Video Coding(HEVC)standard.Therefore,high efficiency video coding has already become current research topic.HEVC inherits hybrid coding architecture of last coding standard,and it uses transformation,quantization,intra/inter prediction and loop filter to code video efficiently,especially the flexible quad-tree partition construction and multi-angle intra prediction method could improve HEVC encoder performance effectively.However,HEVC also uses Rate-Distortion Optimization(RDO)method which calculates all the rate distortion cost values of all prediction modes and then selects the minimum value as the best mode.Moreover,this RDO method has large computation complexity and additional prediction modes also increase the calculation times,which makes HEVC coder computation complexity improve apparently.To address this problem,this paper proposes two methods:(1)An efficient Uniform Scalar Quantizer and Rate Distortion Optimized Quantization(RDOQ)decision algorithm.This method compares quantization level of these two methods and then proposes decision conditions to select quantizer adaptively.First,if the two quantization level values are equal,HEVC would use RDOQ method,otherwise,HEVC uses the decision conditions to decide the best quantization method.The proposed decision algorithm can select the low computation Uniform Scalar Quantizer and low bits rate RDOQ self-adaptively,and reduce the computation complexity of HEVC encoder.The experimental results show that our proposed algorithm can save about 11% coding time compare with original HEVC encoder.(2)An early skip mode detection algorithm.First,the frames can be classified into texture motionless and texture motion areas based on motion vector and residual information.Meanwhile,the skip modes can be detected in motionless areas.Then,the algorithm employs the difference of entropy to propose a distance decision condition and decide the skip modesin texture motion areas finally.If the coding block meets the conditions,it would encode as skip mode.This proposed early detection algorithm omits the RDO process and saves large cost calculation.The experimental results show that this proposed can save about 39% coding time compare with original HEVC encoder.This paper proposed two efficient algorithms to reduce HEVC computation complexity and both realized time saving,which contributes HEVC to real time application.