Super Sparse 3D Object Detection

doi:10.1109/TPAMI.2023.3286409

CORC > 自动化研究所 > 中国科学院自动化研究所 > 智能感知与计算研究中心

	Super Sparse 3D Object Detection
	Fan, Lue 1,2; Yang, Yuxue 1,3; Wang, Feng 5; Wang, Naiyan 5; Zhang, Zhaoxiang 1,3,4
刊名	IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
	2023-10-01
卷号	45 期号:10 页码:12490-12505
关键词	Feature extraction Detectors Point cloud compression Three-dimensional displays Proposals Object detection Laser radar 3D object detection autonomous driving instance segmentation LiDAR point clustering sparse temporal fusion waymo open dataset
ISSN号	0162-8828
DOI	10.1109/TPAMI.2023.3286409
通讯作者	Zhang, Zhaoxiang(zhaoxiang.zhang@ia.ac.cn)
英文摘要	As the perception range of LiDAR expands, LiDAR-based 3D object detection contributes ever-increasingly to the long-range perception in autonomous driving. Mainstream 3D object detectors often build dense feature maps, where the cost is quadratic to the perception range, making them hardly scale up to the long-range settings. To enable efficient long-range detection, we first propose a fully sparse object detector termed FSD. FSD is built upon the general sparse voxel encoder and a novel sparse instance recognition (SIR) module. SIR groups the points into instances and applies highly-efficient instance-wise feature extraction. The instance-wise grouping sidesteps the issue of the center feature missing, which hinders the design of the fully sparse architecture. To further enjoy the benefit of fully sparse characteristic, we leverage temporal information to remove data redundancy and propose a super sparse detector named FSD++. FSD++ first generates residual points, which indicate the point changes between consecutive frames. The residual points, along with a few previous foreground points, form the super sparse input data, greatly reducing data redundancy and computational overhead. We comprehensively analyze our method on the large-scale Waymo Open Dataset, and state-of-the-art performance is reported. To showcase the superiority of our method in long-range detection, we also conduct experiments on Argoverse 2 Dataset, where the perception range (200 m) is much larger thanWaymo Open Dataset (75 m).
资助项目	National Natural Science Foundation of China[2018AAA0100400] ; [61836014] ; [U21B2042] ; [62072457] ; [62006231]
WOS关键词	TIME
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	IEEE COMPUTER SOC
WOS记录号	WOS:001197545600001
资助机构	National Natural Science Foundation of China
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/58101]
专题	自动化研究所_智能感知与计算研究中心
通讯作者	Zhang, Zhaoxiang
作者单位	1.Chinese Acad Sci CASIA, Inst Automat, Natl Lab Pattern Recognit NLPR, Ctr Res Intelligent Percept & Comp CRIPAC, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci UCAS, Sch Future Technol, Beijing 100049, Peoples R China 3.Univ Chinese Acad Sci UCAS, Sch Artificial Intelligence, Beijing 100049, Peoples R China 4.Chinese Acad Sci HKISI CAS, Hong Kong Inst Sci & Innovat, Ctr Artificial Intelligence & Robot, Hong Kong, Peoples R China 5.Tusimple, Beijing 100020, Peoples R China
推荐引用方式 GB/T 7714	Fan, Lue,Yang, Yuxue,Wang, Feng,et al. Super Sparse 3D Object Detection[J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,2023,45(10):12490-12505.
APA	Fan, Lue,Yang, Yuxue,Wang, Feng,Wang, Naiyan,&Zhang, Zhaoxiang.(2023).Super Sparse 3D Object Detection.IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,45(10),12490-12505.
MLA	Fan, Lue,et al."Super Sparse 3D Object Detection".IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 45.10(2023):12490-12505.