自适应分割的视频点云多模式帧间编码方法

陈建; 廖燕俊; 王适; 郑明魁; 苏立超

doi:10.16383/j.aas.c220549

自适应分割的视频点云多模式帧间编码方法

doi: 10.16383/j.aas.c220549

陈建^{1, 2,},
廖燕俊^1,,
王适^2,,
郑明魁^{1, 2,},
苏立超^3,

1.
福州大学先进制造学院泉州 362251
2.
福州大学物理与信息工程学院福州 350116
3.
福州大学计算机与大数据学院/软件学院福州 350116

基金项目: 国家自然科学基金(62001117, 61902071), 福建省自然科学基金(2020J01466), 中国福建光电信息科学与技术创新实验室(闽都创新实验室) (2021ZR151), 超低延时视频编码芯片及其产业化(2020年福建省教育厅产学研专项)资助

详细信息

作者简介:
陈建：福州大学物理与信息工程学院副教授. 主要研究方向为视频编码, 压缩感知, 点云压缩和目标跟踪. E-mail: chenjian-fzu@163.com

廖燕俊：福州大学先进制造学院硕士研究生. 主要研究方向为点云分割和视频点云压缩. E-mail: liao.yanjun@foxmail.com

王适：福州大学物理与信息工程学院硕士研究生. 主要研究方向为多媒体技术. E-mail: wang_kuo@cib.com.cn

郑明魁：福州大学物理与信息工程学院副教授. 主要研究方向为计算机视觉, 点云与视频编码. 本文通信作者. E-mail: zhengmk@fzu.edu.cn

苏立超：福州大学计算机与大数据学院/软件学院讲师. 主要研究方向为多媒体信息安全. E-mail: fzu-slc@fzu.edu.cn

计量
- 文章访问数: 364
- HTML全文浏览量: 109
- PDF下载量: 127
- 被引次数: 0
出版历程
- 收稿日期: 2022-07-05
- 录用日期: 2022-11-29
- 网络出版日期: 2023-07-06
- 刊出日期: 2023-08-21

An Adaptive Segmentation Based Multi-mode Inter-frame Coding Method for Video Point Cloud

CHEN Jian^{1, 2
,},
LIAO Yan-Jun^1
,,
WANG Kuo^2
,,
ZHENG Ming-Kui^{1, 2
,},
SU Li-Chao^3
,

1.
School of Advanced Manufacturing, Fuzhou University, Quanzhou 362251
2.
College of Physics and Information Engineering, Fuzhou University, Fuzhou 350116
3.
College of Computer and Data Science/College of Software, Fuzhou University, Fuzhou 350116

Funds: Supported by National Natural Science Foundation of China (62001117, 61902071), Fujian Natural Science Foundation (2020J01466), Fujian Science & Technology Innovation Laboratory for Optoelectronic Information of China (2021ZR151), and Ultra-low Latency Video Coding Chip and its Industrialization (2020 Special Project of Fujian Provincial Education Department for Industry-University Research)

More Information

Author Bio:
CHEN Jian　Associate professor at the College of Physics and Information Engineering, Fuzhou University. Her research interest covers video coding, compression sensing, point cloud compression and target tracking

LIAO Yan-Jun　Master student at the School of Advanced Manufacturing, Fuzhou University. His research interest covers point cloud segmentation and dynamic point cloud compression

WANG Kuo　Master student at the College of Physics and Information Engineering, Fuzhou University. His main research interest is multimedia technology

ZHENG Ming-Kui　Associate professor at the College of Physics and Information Engineering, Fuzhou University. His research interest covers computer vision, video and point cloud coding. Corresponding author of this paper

SU Li-Chao　Lecturer at the College of Computer and Data Science/College of Software, Fuzhou University. His main research interest is multimedia infomation security

摘要

摘要: 基于视频的点云压缩(Video based point cloud compression, V-PCC)为压缩动态点云提供了高效的解决方案, 但V-PCC从三维到二维的投影使得三维帧间运动的相关性被破坏, 降低了帧间编码性能. 针对这一问题, 提出一种基于V-PCC改进的自适应分割的视频点云多模式帧间编码方法, 并依此设计了一种新型动态点云帧间编码框架. 首先, 为实现更精准的块预测, 提出区域自适应分割的块匹配方法以寻找最佳匹配块; 其次, 为进一步提高帧间编码性能, 提出基于联合属性率失真优化(Rate distortion optimization, RDO)的多模式帧间编码方法, 以更好地提高预测精度和降低码率消耗. 实验结果表明, 提出的改进算法相较于V-PCC实现了−22.57%的BD-BR (Bjontegaard delta bit rate)增益. 该算法特别适用于视频监控和视频会议等帧间变化不大的动态点云场景.
- 点云压缩 /
- 基于视频的点云压缩 /
- 三维帧间编码 /
- 点云分割 /
- 率失真优化
Abstract: Video based point cloud compression (V-PCC) provides an efficient solution for compressing dynamic point clouds, but the projection of V-PCC from 3D to 2D destroys the correlation of 3D inter-frame motion and reduces the performance of inter-frame coding. To solve this problem, we proposes an adaptive segmentation based multi-mode inter-frame coding method for video point cloud to improve V-PCC, and designs a new dynamic point cloud inter-frame encoding framework. Firstly, in order to achieve more accurate block prediction, a block matching method based on adaptive regional segmentation is proposed to find the best matching block; Secondly, in order to further improve the performance of inter coding, a multi-mode inter-frame coding method based on joint attribute rate distortion optimization (RDO) is proposed to increase the prediction accuracy and reduce the bit rate consumption. Experimental results show that the improved algorithm proposed in this paper achieves −22.57% Bjontegaard delta bit rate (BD-BR) gain compared with V-PCC. The algorithm is especially suitable for dynamic point cloud scenes with little change between frames, such as video surveillance and video conference.
- Point cloud compression /
- video-based point cloud compresion (V-PCC) /
- 3D inter-frame coding /
- point cloud segmentation /
- rate distortion optimization (RDO)

HTML全文

图 1 V-PCC编码器框架

Fig. 1 V-PCC encoder diagram

下载: 全尺寸图片幻灯片

图 2 V-PCC从三维到二维投影(属性图)

Fig. 2 V-PCC projection from 3D to 2D (Attribute map)

下载: 全尺寸图片幻灯片

图 3 改进的三维帧间编码框架

Fig. 3 Improved 3D inter-frame coding framework

下载: 全尺寸图片幻灯片

图 4 区域自适应分割块匹配方法示意图

Fig. 4 Schematic diagram of region adaptive segmentation based block matching method

下载: 全尺寸图片幻灯片

图 5 区域自适应分割的块匹配方法分割示例

Fig. 5 Example of block matching method based on adaptive regional segmentation

下载: 全尺寸图片幻灯片

图 6 率失真折线与$\lambda$示意图

Fig. 6 Rate distortion polyline and $\lambda$ diagram

下载: 全尺寸图片幻灯片

图 7 8i动态点云序列示例

Fig. 7 Example of 8i dynamic point cloud sequence

下载: 全尺寸图片幻灯片

图 8 消融实验率失真曲线图

Fig. 8 Rate distortion curve of ablation experiment

下载: 全尺寸图片幻灯片

图 9 Soldier序列率失真曲线图

Fig. 9 Rate distortion curve of Soldier sequence

下载: 全尺寸图片幻灯片

表 1 相对近似块帧间编码比特占用

Table 1 RSB inter coding bits occupancy

编码类型	编码信息	数据类型	比特占用(bits)
位置信息	$(X,Y,Z)_{{\rm{min}},{\rm{max}}}$	Int	96
旋转矩阵	$3\times3$矩阵	Float	288
平移向量	$3\times 1$向量	Float	96
属性偏移	$(R,G,B)$	Int	48
总编码消耗	$R_{\rm{RSB}}$	Int、Float	528

下载: 导出CSV

表 2 提出的改进算法量化参数

Table 2 Quantization parameters of proposed improved algorithm

量化参数	编码模式判定$\Omega$	最小块阈值$N_D$	分割截止阈值$T$
R1	3.5	14000	2.0
R2	3.0	12000	2.2
R3	2.5	10000	2.5
R4	2.0	8000	2.8
R5	1.5	6000	3.0

下载: 导出CSV

表 3 区域自适应分割的块匹配方法性能测试

Table 3 Performance test of block matching method based on adaptive regional segmentation

	Name	D1 (%)	D2 (%)	Luma (%)	Cb (%)	Cr (%)
均匀块分割	Loot	33.93	49.67	4.48	−18.07	−15.08
	Redandblack	13.51	19.42	13.41	−3.01	10.79
	Soldier	−34.75	−34.66	−40.82	−42.91	−44.52
	Queen	−33.64	−33.33	−18.33	−28.58	−12.74
	Longdress	11.97	13.15	27.48	−4.88	11.13
	Average	−1.80	−2.85	−2.76	−18.51	−10.08
改进的自适应块分割	Loot	27.10	39.28	2.22	−24.14	−21.19
	Redandblack	6.03	8.14	16.06	4.14	13.72
	Soldier	−37.69	−37.57	−42.87	−45.10	−46.42
	Queen	−35.72	−33.94	−9.36	−12.93	−14.92
	Longdress	6.74	6.46	4.83	5.02	12.49
	Average	−6.71	−3.53	−5.82	−14.60	−11.26
相对BD-BR增益		−4.91	−6.38	−3.06	3.91	−1.18

下载: 导出CSV

表 4 联合属性率失真优化的多模式帧间编码性能测试

Table 4 Performance test of multi-mode inter-frame coding based on joint attribute rate distortion optimization

	Name	D1 (%)	D2 (%)	Luma (%)	Cb (%)	Cr (%)
完全帧间运动预测	Loot	32.31	42.75	11.16	−19.63	−16.09
	Redandblack	14.16	15.52	7.88	−11.11	6.35
	Soldier	−31.21	−31.21	−37.47	−41.07	−42.48
	Queen	−29.34	−30.23	−12.47	−17.84	−6.77
	Longdress	12.28	14.79	17.74	−13.26	0.79
	Average	−0.36	2.32	−2.63	−20.58	−13.70
改进的多模式帧间运动预测	Loot	−0.06	0.70	10.07	5.44	5.80
	Redandblack	−3.31	−2.09	11.73	1.66	9.45
	Soldier	−35.88	−34.59	−36.31	−36.28	−38.70
	Queen	−39.55	−39.37	−39.16	−44.46	−44.97
	Longdress	−0.64	1.28	6.50	3.10	10.23
	Average	−18.13	−17.09	−9.43	−16.06	−11.64
相对BD-BR增益		−17.77	−19.41	−6.80	4.52	2.06

下载: 导出CSV

表 5 相较于V-PCC相互消融性能(平均BD-BR)

Table 5 Mutual ablation performance compared to V-PCC (Average BD-BR)

应用算法	D1 (%)	D2 (%)	Luma (%)	Cb (%)	Cr (%)
多模式帧间编码	−18.13	−17.09	−9.43	−16.06	−11.64
自适应块分割	−6.71	−3.53	−5.82	−14.60	−11.26
完整算法框架	−22.57	−20.94	−22.01	−23.67	−21.90

下载: 导出CSV

表 6 提出的改进算法相较于V-PCC的性能比较

Table 6 Performance of proposed improved algorithm compared with V-PCC

Name	D1 (%)	D2 (%)	Luma (%)	Cb (%)	Cr (%)
Loot	−2.64	−0.69	−3.57	−4.74	−4.85
Redandblack	−1.65	−0.95	−1.75	−2.01	−2.12
Soldier	−51.53	−46.99	−54.12	−55.65	−55.95
Queen	−56.58	−56.45	−51.62	−55.12	−46.27
Longdress	−0.44	0.40	0.99	−0.80	−0.30
Overall	−22.57	−20.94	−22.01	−23.67	−21.90

下载: 导出CSV

表 7 提出的改进算法同文献[15, 21-22]的性能对比

Table 7 Performance comparisons between the proposed improved algorithm and references [15, 21-22]

Name	D1 (%)	D2 (%)	Luma (%)	Cb (%)	Cr (%)
文献[15]	−21.02	−15.68	−3.54	—	—
文献[21]	−4.70	−4.70	−9.70	−16.30	−14.50
文献[22]	−14.80	−19.10	−2.20	−16.20	−19.90
提出的算法	−22.57	−20.94	−22.01	−23.67	−21.90

下载: 导出CSV

表 8 相较于V-PCC的编码开销与时间复杂度

Table 8 Encoding overhead and time complexity compared to V-PCC

	量化等级	预测编码 (bits)	NSB 编码 (bits)	V-PCC 帧内编码(bits)	相对编码开销(%)	相对编码时间(%)
Queen	R1	5722	33065	97363	39.83	85.01
	R2	6099	48154	140968	38.49	84.74
	R3	6841	78773	225197	38.02	84.99
	R4	6797	149308	392872	39.71	85.77
	R5	7494	292382	686758	43.67	92.08
Soldier	R1	4986	47469	149301	35.13	82.17
	R2	5692	76517	236022	34.83	81.79
	R3	7986	138810	392283	37.42	81.27
	R4	11836	305130	675218	46.94	83.38
	R5	16182	826968	1192585	70.69	94.29
Average		7955	199658	418857	49.56	85.55

下载: 导出CSV

参考文献(28)

[1]	李厚强, 李礼, 李竹. 点云编码综述. 中兴通讯技术, 2021, 27(01):5−9 Li Hou-Qiang, Li Li, Li Zhu. A review of point cloud compression. ZTE Technology Journal, 2021, 27(01):5−9
[2]	Ishikawa Y, Hachiuma R, Ienaga N, Kuno W, Sugiura Y, Saito H. Semantic segmentation of 3D point cloud to virtually manipulate real living space. In: Proceedings of the Asia Pacific Workshop on Mixed and Augmented Reality. Ikoma, Japan: IEEE, 2019. 1−7
[3]	Deng X M, Zhu Y Y, Zhang Y D, Cui Z P, Tan P, Qu W, et al. Weakly supervised learning for single depth-based hand shape recovery. IEEE Transactions on Image Processing, 2021, 30: 532−545 doi: 10.1109/TIP.2020.3037479
[4]	田永林, 沈宇, 李强, 王飞跃. 平行点云:虚实互动的点云生成与三维模型进化方法. 自动化学报, 2020, 46(12): 2572−2582 doi: 10.16383/j.aas.c200800 Tian Yong-Lin, Shen Yu, Li Qiang, Wang Fei-Yue. Parallel point clouds: point clouds generation and 3D model evolution via virtual-real interaction. Acta Automatica Sinica, 2020, 46(12): 2572−2582 doi: 10.16383/j.aas.c200800
[5]	Sun P P, Zhao X M, Xu Z G, Wang R M, Min H G. A 3D LiDAR data-based dedicated road boundary detection algorithm for autonomous vehicles. IEEE Access, 2019, 7: 29623−29638 doi: 10.1109/ACCESS.2019.2902170
[6]	Huang X Y, Wang P, Cheng X J, Zhou D F, Geng Q C, Yang R G. The apolloscape open dataset for autonomous driving and its application. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(10): 2702−2719 doi: 10.1109/TPAMI.2019.2926463
[7]	Gan Z P, Xu H R, He Y R, Cao W, Chen G H. Autonomous landing point retrieval algorithm for UAVs based on 3D environment perception. In: Proceedings of the IEEE 7th International Conference on Virtual Reality. Foshan, China: IEEE, 2021. 104−108
[8]	曹成坤, 张琮毅, 汪国平. 精度可控的字典全局相似性点云压缩. 计算机辅助设计与图形学学报, 2019, 31(6):869−877 Cao Cheng-Kun, Zhang Cong-Yi, Wang Guo-Ping. Precision controllable point clouds compression using global similarity in dictionary. Journal of Computer-Aided Design & Computer Graphics, 2019, 31(6): 869−877
[9]	Maja K, Chou P A, Savill P. 8i voxelized surface light field (8iVSLF) dataset. ISO/IEC JTC1/SC29/WG11 MPEG, input document m42914, Ljubljana, Slovenia, 2018
[10]	Kammerl J, Blodow N, Rusu R B, Gedikli S, Beetz M, Steinbach E. Real-time compression of point cloud streams. In: Proceedings of the IEEE International Conference on Robotics and Automation. Saint Paul, USA: IEEE, 2012. 778−785
[11]	Thanou D, Chou P A, Frossard P. Graph-based compressionof dynamic 3D point cloud sequences. IEEE Transactions on Image Processing, 2016, 25(4): 1765−1778 doi: 10.1109/TIP.2016.2529506
[12]	Queiroz R L, Chou P A. Motion-compensated compression of dynamic voxelized point clouds. IEEE Transactions on Image Processing, 2017, 26(8): 3886−3895 doi: 10.1109/TIP.2017.2707807
[13]	Mekuria R, Blom K, Cesar P. Design, implementation, and evaluation of a point cloud codec for tele-immersive video. IEEE Transactions on Circuits and Systems for Video Technology, 2017, 27(4): 828−842 doi: 10.1109/TCSVT.2016.2543039
[14]	Besl P J, McKay N D. A method for registration of 3-D shapes. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2):239−256 doi: 10.1109/34.121791
[15]	Santos C, Goncalves M, Correa G, Porto M. Block-based inter-frame prediction for dynamic point cloud compression. In: Proceedings of the IEEE International Conference on Image Processing. Anchorage, USA: IEEE, 2021. 3388−3392
[16]	Lasserre S, Llach J, Guede C, Ricard J. Technicolor＇s response to the CfP for point cloud compression. ISO/IEC JTC1/SC29/WG11 MPEG, input document m41822, Macau, China, 2017
[17]	Budagavi M, Faramarzi E, Ho T, Najaf-Zadeh H, Sinharoy I. Samsung＇s response to CfP for point cloud compression (Category2). ISO/IEC JTC1/SC29/WG11 MPEG, input document m41808, Macau, China, 2017
[18]	Schwarz S, Sheikhipour N, Sevom V F, Hannuksela M M. Video coding of dynamic 3D point cloud data. APSIPA Transactions on Signal and Information Processing, DOI: 10.1017/ATSIP.2019.24
[19]	Mammou K, Tourapis A M, SINGER D. Video-based and hierarchical approaches point cloud compression. ISO/IEC JTC1/SC29/WG11 MPEG, input document m41779, Macua, China, 2017
[20]	Graziosi D, Nakagami O, Kuma S, Zaghetto A, Suzuki T, Tabatabai A. An overview of ongoing point cloud compression standardization activities: Video-based (V-PCC) and geometry-based (G-PCC). APSIPA Transactions on Signal and Information Processing, DOI: 10.1017/ATSIP.2020.12
[21]	Li L, Zhu L, Zakharchenko V, Chen J L, Li H Q. Advanced 3D motion prediction for Video-based dynamic point cloud compression. IEEE Transactions on Image Processing, 2020, 29:289−302 doi: 10.1109/TIP.2019.2931621
[22]	Kim J, Im J, Rhyu S, Kim K. 3D motion estimation and compensation method for Video-based point cloud compression. IEEE Access, 2020, 8:83538−83547 doi: 10.1109/ACCESS.2020.2991478
[23]	Costa A, Dricot A, Brites C, Ascenso J, Pereira F. Improved patch packing for the MPEG V-PCC standard. In: Proceedings of the IEEE 21st International Workshop on Multimedia Signal Processing. Kuala Lumpur, Malaysia: IEEE, 2019. 1−6
[24]	Li L, Li Z, Liu S, Li H Q. Occupancy-map-based rate distortion optimization and partition for video-based point cloud compression. IEEE Transactions on Circuits and Systems for Video Technology, 2020, 31(1): 326−338
[25]	Li L, Li Z, Liu S, Li H Q. Efficient projected frame padding for video-based point cloud compression. IEEE Transactions on Multimedia, 2021, 23: 2806−2819 doi: 10.1109/TMM.2020.3016894
[26]	Xu Y Q, Hu W, Wang S S, Zhang X F, et al. Predictive generalized graph fourier transform for attribute compression of dynamic point clouds. IEEE Transactions on Circuits and Systems for Video Technology, 2021, 31(5): 1968−1982 doi: 10.1109/TCSVT.2020.3015901
[27]	Schwarz S, Martin-Cocher G, Flynn D, Budagavi M. Common test conditions for point cloud compression. ISO/IEC JTC1/SC29/WG11 MPEG, input docment w17766, Ljubljana, Slovenia, 2020
[28]	Jang E S, Preda M, Mammou K, Tourapis A M, et al. Video-based point-cloud-compression standard in MPEG: from evidence collection to committee draft. IEEE Signal Processing Magazine, 2019, 36(3): 118−123 doi: 10.1109/MSP.2019.2900721