| With the continuous development of multimedia technology,video application trend of diversification has put forward higher request to video compression technology.Therefore,in April 2010 VCEG and MPGE jointly establish JCT-VC,jointly developed a new generation of video coding standard-H.265 / HEVC,and release standard in 2013.Although H.265 / HEVC than H.264 / AVC,under the same image quality can save 50% of the rate,but for its computational complexity is increased from 2 to 4 times.So how to improve the coding speed of new standard HEVC,reduce the computational complexity has become now the main topics of the researchers.Based on this purpose,this paper respectively from two aspects of inter-frame coding and intra coding of H.265 / HEVC coding optimization,specific work is as follows:Contained within the intra coding,In this paper,based on image gray histogram and spatial correlation.So the whole optimization process is divided into two parts,The first part based on the grey value distribution probability of image gray histogram LCU roughing,using the brightness values are classified in the LCU piece,if the distribution of the two pieces of similar,then is coded reference information can be encoded blocks,namely can preserve 32×32 chunks of code blocks,unnecessary to smaller division.The second part is based on the spatial correlation on the depth of the LCU block coding block decisions,this method USES frame there is strong correlation between adjacent radiating elements in theory the coding unit is divided for reference,studies have shown that frame coding block and its upper part,the left and the upper left cell block partition similarity is higher,therefore,in the current code of CU block division,you can refer to the division of the adjacent CU block it.By these two steps,the frame coding unit can be divided into early termination,save the encoding time,the experimental results show that this method can save 22.84% of the average encoding time and PSNR reduce 0.031 db,this suggests that the visual quality of video not drop too much,and compression ratio is guaranteed,but get reduce encoding time.In terms of inter coding,coding is proposed in this paper in the CU layer content correlation optimization algorithm,through to the threshold set in advance the classified methods of predicting coding unit CU,which can effectively reduce the encoding time.This article puts forward a method based on image content relevance of interframe coding unit depth decision algorithm,starting from the brightness of the image value,by reference to the brightness of the encoded frame between the average difference,by setting the threshold,the early termination of CU element division.In order to reduce the amount of calculation,this paper only select each LCU 21 key pixels are compared,and the calculation of the mean brightness difference in mean when a piece of the brightness of the LCU and its previous frame in the same position of LCU mean brightness difference in the range of the threshold,you can keep the divisions of block 64×64,if is greater than the threshold value and the reference block is divided into 64×64,you can refer to a frame in the same position before the block division,on the other hand,in accordance with the standard of HM traverse division.Experiments show that this algorithm can save 28.959%encoding time,average PSNR changes are relatively small,only 0.0504 db. |
PDF Full Text Request