Researches On Low-complexity Coding Methods For High Efficiency Video Coding

With the rapid development of Internet technology and related multimedia technology,especially the high definition,ultra-high definition and free view video technology,the quality and experience requirements of the video are getting higher and higher.The traditional video coding standard,such as MPEG-4,H.264/AVC(Advanced Video Coding),can no longer meet the requirement of high-definition video coding.To address this problem,the Joint Collaborative Team on Video Coding(JCT-VC)puts forward an efficient solution,namely High Efficiency Video Coding(HEVC)standard,which identified as a new generation of international video coding standard.As compared to its predecessor H.264/AVC,HEVC introduces a lot of new coding techniques and the compression efficiency is almost doubled.Meanwhile,the encoding complexity in HEVC has doubled and redoubled.In order to make the HEVC standard practical and industrialization as soon as possible,a number of solutions to reduce the complexity of HEVC encoders have been proposed by the researches in the academic community.This paper mainly focuses on fast algorithm for HEVC standard,and proposes a series of fast encoding schemes and an encoding complexity control scheme.The main contributions of this paper are:(1)This paper proposed a fast intra coding algorithm for HEVC based on texture property and spatial correlation.The existing intra-frame fast coding schemes only use part of encoding properties,such as the spatial-temporal correlation,texture features or statistical properties of video images.This paper utilizes comprehensively the spatial correlation,texture features and statistical properties of video images,an intra-frame hierarchical fast coding algorithm is proposed.The experimental results show that,compared with the original HEVC encoding scheme,the proposed method reduces encoding time by an average 41.81% with negligible rate distortion performance loss,which is beneficial to the real-time applications of high-definition video coding.(2)This paper proposed a fast inter-frame coding method for HEVC based on motion characteristic and quad-tree models.In the process of HEVC video coding,the most complex process is inter prediction and motion estimation.Therefore,in this paper,we choose two angles from the inter-frame CU partition and the inter mode selection to reduce the computational complexity of inter coding.The existing fast inter coding schemes mainly focus on texture property,motion information,spatial-temporal correlations or statistical information about the CU size and prediction modes to speed up HEVC encoders.However,they do not make full use of these properties,and most of them cannot consider the motion and edge information of different CU depth levels.We proposed an effective fast inter coding algorithm,considering the motion and edge information of each CU depth level.Experimental results show that the proposed algorithm,compared with the latest HEVC encoding scheme,reduces encoding time by 47.5% on average with a Bjonteggard delta bit rate increase of only 1.6% for various test sequences under random access condition.Compared with the state-of-the-art algorithm,the proposed method can save more encoding time while maintaining comparative rate distortion performance.(3)This paper proposed an encoding complexity control algorithm of HEVC for powerconstrained devices.Through intra-prediction and inter-prediction coding analysis,we can see that there is a trade-off between rate distortion performance and coding quality in HEVC fast coding process.The goal of encoding complexity control is to allocate encoding complexity to satisfy the requirements of different HEVC application while maintaining good rate distortion performance.In order to manipulate the encoding complexity of HEVC precisely,this paper explores the statistical characteristics of the coding structure and the relationship between the maximum coding depth and the coding complexity of the HEVC,and allocates the encoding complexity in terms of the GOP,frame and LCTU layers.Experimental results show that the proposed algorithm can reach target encoding complexity of up to 60% with acceptable rate distortion performance reduction under low delay P-main condition,and maintain high control accuracy.