Spatio-Temporal Context-Guided Algorithm for Lossless Point Cloud Geometry Compression

doi:10.12142/ZTECOM.202304003

ZTE Communications ›› 2023, Vol. 21 ›› Issue (4): 17-28.DOI: 10.12142/ZTECOM.202304003

收稿日期:2023-09-11 出版日期:2023-12-07 发布日期:2023-12-07

Spatio-Temporal Context-Guided Algorithm for Lossless Point Cloud Geometry Compression

ZHANG Huiran¹^,², DONG Zhen³(), WANG Mingsheng¹^,²

^1.Guangzhou Urban Planning and Design Survey Research Institute, Guangzhou 510060, China
^2.Guangdong Enterprise Key Laboratory for Urban Sensing, Monitoring and Early Warning, Guangzhou 510060, China
^3.State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing, Wuhan University, Wuhan 430079, China

Received:2023-09-11 Online:2023-12-07 Published:2023-12-07
About author:ZHANG Huiran received her BE and ME degrees in School of Geodesy and Geomatics and State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing, both from Wuhan University, China in 2020 and 2023, respectively. She is currently the surveyor of Guangzhou Urban Planning and Design Survey Research Institute, China. Her research interests include point cloud data processing and compression. She participated in several projects related to the field of remote sensing and published one paper in Geomatics and Information Science of Wuhan University.
DONG Zhen (dongzhenwhu@whu.edu.cn) received his BE and PhD degrees in remote sensing and photogrammetry from Wuhan University, China in 2011 and 2018, respectively. He is a professor with the State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), Wuhan University. His research interests include 3D reconstruction, scene understanding, point cloud processing as well as their applications in intelligent transportation system, digital twin cities, urban sustainable development and robotics. He received over 10 honors from various national and international competitions and published around 60 papers in various journals and conferences.
WANG Mingsheng received his BE degree in College of Computer Science and Technology from Jilin University, China in 2001, and ME degree in School of Computer Science and Engineering from South China University of Technology, China in 2004. He is currently a senior engineer with Guangzhou Urban Planning & Design Survey Research Institute, China. His research interests include computer applications and software, physiography, and surveying. He received over 20 honors from various national competitions and published around 50 papers in various journals and conferences.

摘要/Abstract

Abstract:

Point cloud compression is critical to deploy 3D representation of the physical world such as 3D immersive telepresence, autonomous driving, and cultural heritage preservation. However, point cloud data are distributed irregularly and discontinuously in spatial and temporal domains, where redundant unoccupied voxels and weak correlations in 3D space make achieving efficient compression a challenging problem. In this paper, we propose a spatio-temporal context-guided algorithm for lossless point cloud geometry compression. The proposed scheme starts with dividing the point cloud into sliced layers of unit thickness along the longest axis. Then, it introduces a prediction method where both intra-frame and inter-frame point clouds are available, by determining correspondences between adjacent layers and estimating the shortest path using the travelling salesman algorithm. Finally, the few prediction residual is efficiently compressed with optimal context-guided and adaptive fast-mode arithmetic coding techniques. Experiments prove that the proposed method can effectively achieve low bit rate lossless compression of point cloud geometric information, and is suitable for 3D point cloud compression applicable to various types of scenes.

Key words: point cloud geometry compression, single-frame point clouds, multi-frame point clouds, predictive coding, arithmetic coding

. [J]. ZTE Communications, 2023, 21(4): 17-28.

ZHANG Huiran, DONG Zhen, WANG Mingsheng. Spatio-Temporal Context-Guided Algorithm for Lossless Point Cloud Geometry Compression[J]. ZTE Communications, 2023, 21(4): 17-28.

图/表 9

参考文献 43

1	MI X X, YANG B S, DONG Z, et al. Automated 3D road boundary extraction and vectorization using MLS point clouds [J]. IEEE transactions on intelligent transportation systems, 2022, 23(6): 5287–5297. DOI: 10.1109/TITS.2021.3052882
2	DONG Z, LIANG F X, YANG B S, et al. Registration of large-scale terrestrial laser scanner point clouds: a review and benchmark [J]. ISPRS journal of photogrammetry and remote sensing, 2020, 163: 327–342. DOI: 10.1016/j.isprsjprs.2020.03.013
3	GRAZIOSI D, NAKAGAMI O, KUMA S, et al. An overview of ongoing point cloud compression standardization activities: video-based (V-PCC) and geometry-based (G-PCC) [J]. APSIPA transactions on signal and information processing, 2020, 9: e13
4	DE QUEIROZ R L, CHOU P A. Compression of 3D point clouds using a region-adaptive hierarchical transform [J]. IEEE transactions on image processing, 2016, 25(8): 3947–3956. DOI: 10.1109/TIP.2016.2575005
5	BLETTERER A, PAYAN F, ANTONINI M, et al. Point cloud compression using depth maps [J]. Electronic imaging, 2016, 2016(21):1–6
6	MEKURIA R, BLOM K, Design CESAR P., implementation, and evaluation of a point cloud codec for tele-immersive video [J]. IEEE transactions on circuits and systems for video technology, 2017, 27(4): 828–842. DOI: 10.1109/TCSVT.2016.2543039
7	DE QUEIROZ R L, CHOU P A. Motion-compensated compression of dynamic voxelized point clouds [J]. IEEE transactions on image processing, 2017, 26(8): 3886–3895. DOI: 10.1109/TIP.2017.2707807
8	CAO C, PREDA M, ZAHARIA T. 3D point cloud compression: a survey [C]//The 24th International Conference on 3D Web Technology. ACM, 2019: 1–9. DOI: 10.1145/3329714.3338130
9	GRAZIOSI D, NAKAGAMI O, KUMA S, et al. An overview of ongoing point cloud compression standardization activities: video-based (V-PCC) and geometry-based (G-PCC) [J]. APSIPA transactions on signal and information processing, 2020, 9(1): e13. DOI: 10.1017/atsip.2020.12
10	HUANG Y, PENG J L, KUO C J, et al. Octree-based progressive geometry coding of point clouds [C]//The 3rd Eurographics/IEEE VGTC Conference on Point-Based Graphics. IEEE, 2016: 103–110
11	FAN Y X, HUANG Y, PENG J L. Point cloud compression based on hierarchical point clustering [C]//Asia-Pacific Signal and Information Processing Association Annual Summit and Conference. IEEE, 2014: 1–7. DOI: 10.1109/APSIPA.2013.6694334
12	DRICOT A, ASCENSO J. Adaptive multi-level triangle soup for geometry-based point cloud coding [C]//The 21st International Workshop on Multimedia Signal Processing (MMSP). IEEE, 2019: 1–6. DOI: 10.1109/MMSP.2019.8901791
13	HE C, RAN L Q, WANG L, et al. Point set surface compression based on shape pattern analysis [J]. Multimedia tools and applications, 2017, 76(20): 20545–20565. DOI: 10.1007/s11042-016-3991-0
14	IMDAD U, ASIF M, AHMAD M, et al. Three dimensional point cloud compression and decompression using polynomials of degree one [J]. Symmetry, 2019, 11(2): 209. DOI: 10.3390/sym11020209
15	SUN X B, MA H, SUN Y X, et al. A novel point cloud compression algorithm based on clustering [J]. IEEE robotics and automation letters, 2019, 4(2): 2132–2139. DOI: 10.1109/LRA.2019.2900747
16	DE OLIVEIRA RENTE P, BRITES C, ASCENSO J, et al. Graph-based static 3D point clouds geometry coding [J]. IEEE transactions on multimedia, 2019, 21(2): 284–299. DOI: 10.1109/TMM.2018.2859591
17	ISO. Geometry-based point cloud compression (G-PCC): [S]. 2021
18	DRICOT A, ASCENSO J. Hybrid octree-plane point cloud geometry coding [C]//The 27th European Signal Processing Conference (EUSIPCO). IEEE, 2019: 1–5
19	ZHANG X, GAO W, LIU S. Implicit geometry partition for point cloud compression [C]//Proceedings of 2020 Data Compression Conference (DCC). IEEE, 2020: 73–82. DOI: 10.1109/DCC47342.2020.00015
20	QUACH M, VALENZISE G, DUFAUX F. Learning convolutional transforms for lossy point cloud geometry compression [C]//The 2019 IEEE International Conference on Image Processing (ICIP). IEEE, 2019: 4320–4324. DOI: 10.1109/ICIP.2019.8803413
21	HUANG T X, LIU Y. 3D point cloud geometry compression on deep learning [C]//The 27th ACM International Conference on Multimedia. ACM, 2019: 890–898. DOI: 10.1145/3343031.3351061
22	GUARDA A F R, RODRIGUES N M M, PEREIRA F. Point cloud coding: Adopting a deep learning-based approach [C]//Picture Coding Symposium (PCS). IEEE, 2020: 1–5. DOI: 10.1109/PCS48520.2019.8954537
23	WANG J Q, ZHU H, MA Z, et al. Learned point cloud geometry compression [EB/OL]. [2023-09-01].
24	AINALA K, MEKURIA R N, KHATHARIYA B, et al. An improved enhancement layer for octree based point cloud compression with plane projection approximation [C]//SPIE Optical Engineering+Applications. SPIE, 2016: 223–231. DOI: 10.1117/12.2237753
25	SCHWARZ S, HANNUKSELA M M, FAKOUR-SEVOM V, et al. 2D video coding of volumetric video data [C]//Picture Coding Symposium (PCS). IEEE, 2018: 61–65. DOI: 10.1109/PCS.2018.8456265
26	FAKOUR SEVOM V, SCHWARZ S, GABBOUJ M. Geometry-guided 3D data interpolation for projection-based dynamic point cloud coding [C]//The 7th European Workshop on Visual Information Processing (EUVIP). IEEE, 2019: 1–6. DOI: 10.1109/EUVIP.2018.8611760
27	KATHARIYA B, LI L, LI Z, et al. Lossless dynamic point cloud geometry compression with inter compensation and traveling salesman prediction [C]//Data Compression Conference. IEEE, 2018: 414. DOI: 10.1109/DCC.2018.00067
28	ISO. Visual volumetric video-based coding (V3C) and video-based point cloud compression: [S]. 2021
29	PARK J, LEE J, PARK S, et al. Projection-based occupancy map coding for 3D point cloud compression [J]. IEIE transactions on smart processing & computing, 2020, 9(4): 293–297. DOI: 10.5573/ieiespc.2020.9.4.293
30	COSTA A, DRICOT A, BRITES C, et al. Improved patch packing for the MPEG V-PCC standard [C]//IEEE 21st International Workshop on Multimedia Signal Processing (MMSP). IEEE, 2019: 1–6. DOI: 10.1109/MMSP.2019.8901690
31	KAMMERL J, BLODOW N, RUSU R B, et al. Real-time compression of point cloud streams [C]//Proceedings of 2012 IEEE International Conference on Robotics and Automation. IEEE, 2012: 778–785. DOI: 10.1109/ICRA.2012.6224647
32	PCL. Point cloud library. [EB/OL]. [2023-09-01].
33	THANOU D, CHOU P A, FROSSARD P. Graph-based compression of dynamic 3D point cloud sequences [J]. IEEE transactions on image processing, 2016, 25(4): 1765–1778. DOI: 10.1109/TIP.2016.2529506
34	LI L, LI Z, ZAKHARCHENKO V, et al. Advanced 3D motion prediction for video based point cloud attributes compression [C]//Data Compression Conference (DCC). IEEE, 2019: 498–507. DOI: 10.1109/DCC.2019.00058
35	ZHAO L L, MA K K, LIN X H, et al. Real-time LiDAR point cloud compression using Bi-directional prediction and range-adaptive floating-point coding [J]. IEEE transactions on broadcasting, 2022, 68(3): 620–635. DOI: 10.1109/TBC.2022.3162406
36	LIN J P, LIU D, LI H Q, et al. M-LVC: Multiple frames prediction for learned video compression [C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE, 2020: 3543–3551. DOI: 10.1109/CVPR42600.2020.00360
37	YANG R, MENTZER F, VAN GOOL L, et al. Learning for video compression with hierarchical quality and recurrent enhancement [C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE, 2020: 6627–6636. DOI: 10.1109/CVPR42600.2020.00666
38	KAYA E C, TABUS I. Lossless compression of point cloud sequences using sequence optimized CNN models [J]. IEEE access, 2022, 10: 83678–83691. DOI: 10.1109/ACCESS.2022.3197295
39	DING S, MANNAN M A, POO A N. Oriented bounding box and octree based global interference detection in 5-axis machining of free-form surfaces [J]. Computer-aided design, 2004, 36(13): 1281-1294
40	ALEXIOU E, VIOLA I, BORGES T M, et al. A comprehensive study of the rate-distortion performance in MPEG point cloud compression [J]. APSIPA transactions on signal and information processing, 2019, 8: e27. doi:10.1017/ATSIP.2019.20
41	PEIXOTO E. Intra-frame compression of point cloud geometry using dyadic decomposition [J]. IEEE signal processing letters, 2020, 27: 246–250. DOI: 10.1109/LSP.2020.2965322
42	RAMALHO E, PEIXOTO E, MEDEIROS E. Silhouette 4D with context selection: lossless geometry compression of dynamic point clouds [J]. IEEE signal processing letters, 2021, 28: 1660–1664. DOI: 10.1109/lsp.2021.3102525
43	ISO. Common test conditions for G-PCC document N00106: ISO/IEC [S]. 2021

Point Cloud Data	BPP/bit				Gains
Point Cloud Data	Single↓	G-PCC↓	PCL-PCC↓	S3D↓	G-PCC/%	PCL-PCC/%	S3D/%
Andrew_vox09	1.118 83	1.135 068	2.074 226	1.12	-1.45	-85.39	-0.10
Andrew_vox10	1.010 745	1.104 986	1.952 745	-	-9.32	-93.20	-
David_vox09	1.058 42	1.114 673	2.105 917	1.06	-5.31	-98.97	-0.15
David_vox10	1.028 09	1.090 388	1.974 752	-	-6.06	-92.08	-
Ricardo_vox09	1.037 76	1.081 282	2.046 144	1.04	-4.19	-97.17	-0.22
Ricardo_vox10	0.965 985	1.068 567	1.944 874	-	-10.62	-101.34	-
Sarah_vox09	1.063 19	1.107 865	2.101 666	1.07	-4.20	-97.68	-0.64
Sarah_vox10	1.012 36	1.065 947	1.978 308	-	-5.29	-95.42	-
Longdress_vox10	0.945 35	1.025 244	2.347 862	0.95	-8.45	-148.36	-0.49
Loot_vox10	0.909 825	0.945 36	2.314 874	0.91	-3.91	-154.43	-0.02
Redandblack_vox10	1.014 15	1.082 107	2.400 688	1.03	-6.70	-136.72	-1.56
Soldier_vox10	0.958 515	1.032 572	2.423 025	0.96	-7.73	-152.79	-0.15
Facade 00009 vox12	6.941 5	7.243 8	9.349 4	-	-4.35	-34.69	-
Facade_00015_vox14	7.937 2	8.638 5	10.030 5	-	-8.84	-26.37	-
Arco_Valentino_ Dense_vox12	9.077 9	9.826 4	11.251 4	-	-8.25	-23.94	-
Palazzo_Carignano_ Dense_vox14	7.647 5	8.164 4	9.943 4	-	-6.76	-30.02	-

Point Cloud Data	BPP/bit				Gains
Point Cloud Data	Single↓	G-PCC↓	PCL-PCC↓	S3D↓	G-PCC/%	PCL-PCC/%	S3D/%
Andrew_vox09	1.118 83	1.135 068	2.074 226	1.12	-1.45	-85.39	-0.10
Andrew_vox10	1.010 745	1.104 986	1.952 745	-	-9.32	-93.20	-
David_vox09	1.058 42	1.114 673	2.105 917	1.06	-5.31	-98.97	-0.15
David_vox10	1.028 09	1.090 388	1.974 752	-	-6.06	-92.08	-
Ricardo_vox09	1.037 76	1.081 282	2.046 144	1.04	-4.19	-97.17	-0.22
Ricardo_vox10	0.965 985	1.068 567	1.944 874	-	-10.62	-101.34	-
Sarah_vox09	1.063 19	1.107 865	2.101 666	1.07	-4.20	-97.68	-0.64
Sarah_vox10	1.012 36	1.065 947	1.978 308	-	-5.29	-95.42	-
Longdress_vox10	0.945 35	1.025 244	2.347 862	0.95	-8.45	-148.36	-0.49
Loot_vox10	0.909 825	0.945 36	2.314 874	0.91	-3.91	-154.43	-0.02
Redandblack_vox10	1.014 15	1.082 107	2.400 688	1.03	-6.70	-136.72	-1.56
Soldier_vox10	0.958 515	1.032 572	2.423 025	0.96	-7.73	-152.79	-0.15
Facade 00009 vox12	6.941 5	7.243 8	9.349 4	-	-4.35	-34.69	-
Facade_00015_vox14	7.937 2	8.638 5	10.030 5	-	-8.84	-26.37	-
Arco_Valentino_ Dense_vox12	9.077 9	9.826 4	11.251 4	-	-8.25	-23.94	-
Palazzo_Carignano_ Dense_vox14	7.647 5	8.164 4	9.943 4	-	-6.76	-30.02	-

Point Cloud Data	Average BPP/bit				Average Gains
Point Cloud Data	Single↓	G-PCC↓	PCL-PCC↓	S3D↓	G-PCC	PCL-PCC	S3D
Microsoft voxelized upper bodies	1.036 923	1.096 097	2.022 329	1.072 5	-5.71%	-95.03%	-3.43%
8i voxelized full bodies	0.956 96	1.021 321	2.371 612	0.962 5	-6.73%	-147.83%	-0.58%
MPEG Facade and architecture	1.158 62	1.198 392	2.336 034	-	-3.43%	-101.62%	-

Point Cloud Data	Average BPP/bit				Average Gains
Point Cloud Data	Single↓	G-PCC↓	PCL-PCC↓	S3D↓	G-PCC	PCL-PCC	S3D
Microsoft voxelized upper bodies	1.036 923	1.096 097	2.022 329	1.072 5	-5.71%	-95.03%	-3.43%
8i voxelized full bodies	0.956 96	1.021 321	2.371 612	0.962 5	-6.73%	-147.83%	-0.58%
MPEG Facade and architecture	1.158 62	1.198 392	2.336 034	-	-3.43%	-101.62%	-

Point Cloud Sequences	BPP/bit					Gains
Point Cloud Sequences	Multiple↓	G-PCC↓	InterEM↓	PCL-PCC↓	S4D↓	G-PCC/%	InterEM/%	PCL-PCC/%	S4D/%
Andrew_vox09	1.072 25	1.135 068	-	2.074 226	1.08	-5.86	-	-93.45	-0.72
Andrew_vox10	0.972 24	1.104 986	-	1.952 745	-	-13.65	-	-100.85	-
David_vox09	1.046 565	1.114 673	-	2.105 917	1.05	-6.51	-	-101.22	-0.33
David_vox10	1.020 547	1.090 388	-	1.974 752	-	-6.84	-	-93.50	-
Ricardo_vox09	0.982 66	1.081 282	-	2.046 144	1.02	-10.04	-	-108.23	-3.80
Ricardo_vox10	0.954 235	1.068 567	-	1.944 874	-	-11.98	-	-103.81	-
Sarah_vox09	1.028 745	1.107 865	-	2.101 666	1.04	-7.69	-	-104.29	-1.09
Sarah_vox10	1.008 465	1.065 947	-	1.978 308	-	-5.70	-	-96.17	-
Longdress_vox10	0.896 585	1.025 244	1.056 275	2.347 862	0.95	-14.35	-17.81	-161.87	-5.96
Loot_vox10	0.861 815	0.945 36	1.009 412	2.314 874	0.89	-9.69	-17.13	-168.60	-3.27
Redandblack_vox10	0.970 43	1.082 107	1.140 317	2.400 688	1.01	-11.51	-17.51	-147.38	-4.08
Soldier_vox10	0.704 24	1.032 572	1.070 037	2.423 025	0.79	-46.62	-51.94	-244.06	-12.18

Spatio-Temporal Context-Guided Algorithm for Lossless Point Cloud Geometry Compression

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 43

相关文章 0

编辑推荐

Metrics

Encoding Time/s
Point cloud data	Multiple	Single	S4D	S3D	G-PCC	InterEM	PCL-PCC
Microsoft voxelized upper bodies	52.1	64.2	806.03	489.72	3.813	-	2.235
8i voxelized full bodies	56.7	66.9	904.67	640.85	7.105	4.708	3.549
MPEG facade and architecture	-	111.2	-	-	15.37	-	22.4
Decoding Time/s
Point cloud data	Multiple	Single	S4D	S3D	G-PCC	InterEM	PCL-PCC
Microsoft voxelized upper bodies	13.7	14.4	117.4	74.03	1.031	-	0.809
8i voxelized full bodies	16.3	17.1	194.25	113.95	1.376	4.10	0.922
MPEG facade and architecture	-	22.4	-	-	2.703	-	7.74