-
摘要:
合并模式通过共享邻域块的运动矢量（MV）来节省编码运动信息比特数，有效提升了编码器率失真性能。然而，当前合并模式中的运动补偿预测（MCP）不够准确。为此，分析了合并模式中的预测残差分布特点，并提出了一种基于曼哈顿距离的加权预测算法作为合并模式的附加候选项。首先，采用邻域合并候选项的运动矢量进行运动补偿预测得到多个预测块；然后，根据候选块位置与像素点的曼哈顿距离对获得的多个预测块进行加权平均得到附加候选项；最后，通过率失真优化（RDO）从附加候选项和原有候选项中选择出最佳的合并模式。实验结果显示：在联合探索测试模型JEM 7.0平台上，所提算法在不同的编码器配置下均获得了率失真性能的提升，其中低延迟P帧下达到了平均1.34%的比特率节省。
-
关键词:
- 视频编码 /
- 率失真优化(RDO) /
- 合并模式 /
- 运动补偿 /
- 加权预测
Abstract:Merge mode saves the number of bits required to encode motion information by sharing the motion vector (MV) in neighboring blocks, which improves the rate-distortion performance of encoders effectively. However, motion compensation prediction (MCP) is not accurate enough in the merge mode currently. Therefore, this paper analyses the characteristics of residual distribution after MCP in the merge mode, and presents a Manhattan distance based weighted prediction method as an additional candidate for the merge mode. First, several predicted blocks are obtained by MCP with motion vectors in neighboring candidates. Second, the additional candidate is obtained by a weighted average method according to Manhattan distances from the neighboring candidate to the pixel points in the predicted blocks obtained. Finally, the best merge mode is selected by rate distortion optimization (RDO) among the additional candidate and the original candidates. The experimental results show that, on the joint exploration test model 7.0 (JEM 7.0), the proposed method achieves rate distortion performance improvement under the different configurations of encoder, where a bitrate saving of 1.34% on average is obtained under the configuration of low delay P frame.
-
图 4 不同位置合并候选项进行运动补偿预测后的残差幅度统计分布
Figure 4. Statistical distribution for residual magnitudes after MCP of merge candidates at different locations
表 1 不同位置候选块对应的权重设置
Table 1. Weight setting for candidates at different locations
候选块的位置 Wi(x, y) A1 W-x+y B1 H-y+x B0 H-y+x A0 W-x+y B2 W-x+H-y+1 TB x+y+1 
下载: 导出CSV
表 2 三种编码器配置下本文算法相比于JEM 7.0的BDBR
Table 2. BDBR for proposed algorithm compared with JEM 7.0 under three configurations of encoder
% Class 测试序列 LDP LDB RA Y U V Y U V Y U V B Kimono -0.32 -0.29 -0.56 -0.07 -0.10 0.02 -0.07 0.11 0.04 ParkScene -0.36 -1.35 -0.83 0.08 0.23 0.30 -0.03 -0.39 -0.42 Cactus -1.66 -3.01 -2.57 -0.17 -0.37 -0.90 -0.26 -0.64 -0.46 BasketballDrive -1.18 -2.20 -2.27 -0.27 -0.77 -0.52 -0.15 -0.38 -0.02 BQTerrace -6.98 -5.93 -5.02 -0.40 -0.76 -0.57 -0.61 -0.71 -0.94 C BasketballDrill -1.58 -2.12 -2.69 -0.05 0.02 -0.17 -0.21 -0.02 -0.13 BQMall -0.56 -0.86 -0.66 -0.01 -0.79 -0.24 -0.08 -0.40 0.15 PartyScene -1.59 -1.66 -1.17 -0.33 -0.43 -0.19 -0.24 0 -0.16 RaceHorses -0.89 -1.87 -2.04 -0.08 -0.33 -0.54 -0.13 -0.28 -0.24 D BasketballPass -0.29 -0.53 -0.86 -0.01 -1.39 -0.60 -0.08 -0.26 -0.02 BQSquare -3.37 -5.00 -3.32 -0.36 -0.20 -1.26 -0.27 -0.03 0 BlowingBubbles -0.84 -0.94 -1.24 -0.23 -0.58 -0.67 -0.04 -0.01 -0.22 RaceHorses -0.12 -0.12 0.08 -0.01 -0.36 -0.45 0.02 -0.21 -0.23 E FourPeople -1.27 -1.17 -0.67 -0.29 -1.60 -2.66 — — — Johnny -2.59 -3.99 -3.59 -0.03 -0.59 -0.52 — — — KristenAndSara -1.10 -3.75 -3.70 0.09 -0.67 0.09 — — — F BasketballDrillText -1.31 -0.98 -1.19 -0.15 -0.40 -0.29 -0.23 -0.04 -0.35 ChinaSpeed -0.29 -1.37 -1.02 -0.08 -0.27 -0.24 -0.09 -0.16 -0.25 SlideEditing -0.28 -0.15 -0.18 -0.02 -0.31 -0.61 0.04 0.03 0.01 SlideShow -0.16 -0.89 0.46 0.49 1.84 0.63 -0.30 -0.41 -0.19 平均 -1.34 -1.91 -1.65 -0.10 -0.39 -0.47 -0.16 -0.22 -0.20
下载: 导出CSV
-
[1] 马思伟.AVS视频编码标准技术回顾及最新进展[J].计算机研究与发展, 2015, 52(1):27-37. http://d.old.wanfangdata.com.cn/Periodical/jsjyjyfz201501004MA S W.History and recent developments of AVS video coding standards[J].Journal of Computer Research and Development, 2015, 52(1):27-37(in Chinese). http://d.old.wanfangdata.com.cn/Periodical/jsjyjyfz201501004 [2] SULLIVAN G J, OHM J R, HAN W J, et al.Overview of the high efficiency video coding(HEVC) standard[J].IEEE Transactions on Circuits and Systems for Video Technology, 2012, 22(12):1649-1668. doi: 10.1109/TCSVT.2012.2221191 [3] GAO Y, ZHU C, LI S, et al.Temporal dependent rate-distortion optimization for low-delay hierarchical video coding[J].IEEE Transactions on Image Processing, 2017, 26(9):4457-4470. doi: 10.1109/TIP.2017.2713598 [4] LI S, ZHU C, GAO Y B, et al.Lagrangian multiplier adaptation for rate-distortion optimization with inter-frame dependency[J].IEEE Transactions on Circuits and Systems for Video Technology, 2016, 26(1):117-129. doi: 10.1109/TCSVT.2015.2450131 [5] LI S, ZHU C, GAO Y, et al.Inter-frame dependent rate-distortion optimization using Lagrangian multiplier adaption[C]//Proceedings of IEEE International Conference on Multimedia and Expo(ICME).Piscataway, NJ: IEEE Press, 2015: 1-6. [6] HELLE P, OUDIN S, BROSS B, et al.Block merging for quadtree-based partitioning in HEVC[J].IEEE Transactions on Circuits and Systems for Video Technology, 2012, 22(12):1720-1731. doi: 10.1109/TCSVT.2012.2223051 [7] PAN Z Q, KWONG S, SUN M T, et al.Early merge mode decision based on motion estimation and hierarchical depth correlation for HEVC[J].IEEE Transactions on Broadcasting, 2014, 60(2):405-412. doi: 10.1109/TBC.2014.2321682 [8] TARIQ J, KWONG S, YUAN H.Spatial/temporal motion consistency based merge mode early decision for HEVC[J].Journal of Visual Communication and Image Representation, 2017, 44:198-213. doi: 10.1016/j.jvcir.2017.01.029 [9] 蒋洁, 叶德周, 潘勉.一种基于自适应阈值的Merge模式快速选择方法[J].光电子·激光, 2016, 27(9):980-986. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=gdzjg201609013JIANG J, YE D Z, PAN M.A fast candidate selection method for Merge mode based on adaptive threshold[J].Journal of Optoelectronics·Laser, 2016, 27(9):980-986(in Chinese). http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=gdzjg201609013 [10] KIM T S, RHEE C E, LEE H J.Merge mode estimation for a hardware-based HEVC encoder[J].IEEE Transactions on Circuits and Systems for Video Technology, 2016, 26(1):195-209. doi: 10.1109/TCSVT.2015.2496820 [11] SONG Y X, JIA K B.Early merge mode decision for texture coding in 3D-HEVC[J].Journal of Visual Communication and Image Representation, 2015, 33:60-68. doi: 10.1016/j.jvcir.2015.07.001 [12] HEO Y S, BANG G, PARK G H.Adaptive merge list construction for 3D-HEVC fast encoder[J].Electronics Letters, 2016, 52(8):604-690. doi: 10.1049/el.2015.4323 [13] 宋雨新, 贾克斌, 吴强.3D-HEVC合并模式快速判决方法研究[J].信号处理, 2016, 32(1):46-55. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=xhcl201601007SONG Y X, JIA K B, WU Q.Research on fast merge mode decision method for 3D-HEVC[J].Journal of Signal Processing, 2016, 32(1):46-55(in Chinese). http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=xhcl201601007 [14] WU J F, GUO B L, HOU J, et al.Fast CU encoding schemes based on merge mode and motion estimation for HEVC inter prediction[J].KSII Transactions on Internet and Information Systems, 2016, 10(3):1195-1211. http://cn.bing.com/academic/profile?id=cc02e852a11680f79c55a31e336390c6&encoded=0&v=paper_preview&mkt=zh-cn [15] CHENG Z X, SUN H M, ZHOU D J, et al.Accelerating HEVC inter prediction with improved merge mode handling[J].IEICE Transactions on Fundamentals of Electronics Communications and Computer Sciences, 2017, 100(2):546-554. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=J-STAGE_660436 [16] VANNE J, VIITANEN M, HAMALAINEN T D.Efficient mode decision schemes for HEVC inter prediction[J].IEEE Transactions on Circuits and Systems for Video Technology, 2014, 24(9):1579-1593. doi: 10.1109/TCSVT.2014.2308453 [17] ZHANG N, FAN X P, ZHAO D B, et al.Merge mode for deformable block motion information derivation[J].IEEE Transactions on Circuits and Systems for Video Technology, 2017, 27(11):2437-2449. doi: 10.1109/TCSVT.2016.2589818 [18] CHEN J, ALSHINA E, SULLIVAN G J, et al.Algorithm description of joint exploration test model 7(JEM 7)[C]//Joint Video Exploration Team(JVET) of ITU-T SG16WP3 and ISO/IEC JTC 1/SC 29/WG 117th Meeting, 2017: 10-15. [19] SUEHRING K, LI X.JVET common test conditions and software reference configurations[C]//Joint Video Exploration Team (JVET) of ITU-T SG16WP3 and ISO/IEC JTC 1/SC 29/WG112nd Meeting, 2016: 1-4. -
下载:
点击查看大图
计量
- 文章访问数: 733
- HTML全文浏览量: 72
- PDF下载量: 298
- 被引次数: 0
