Self-Supervised Point Cloud Completion on Real Traffic Scenes Via Scene-Concerned Bottom-Up Mechanism

doi:10.1109/ICME52920.2022.9860015

	Self-Supervised Point Cloud Completion on Real Traffic Scenes Via Scene-Concerned Bottom-Up Mechanism
	Yiming Ren1 ; Peishan Cong1 ; Xinge Zhu 2; Yuexin Ma1
	2022
会议录名称	PROCEEDINGS - IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO
ISSN	1945-7871
卷号	2022-July
发表状态	已发表
DOI	10.1109/ICME52920.2022.9860015
摘要	Real scans always miss partial geometries of objects due to the self-occlusions, external-occlusions, and limited sensor resolutions. Point cloud completion aims to refer the complete shapes for incomplete 3D scans of objects. Current deep learning-based approaches rely on large-scale complete shapes in the training process, which are usually obtained from synthetic datasets. It is not applicable for real-world scans due to the domain gap. In this paper, we propose a self-supervised point cloud completion method (TraPCC) for vehicles in real traffic scenes without any complete data. Based on the symmetry and similarity of vehicles, we make use of consecutive point cloud frames to construct vehicle memory bank as reference. We design a bottom-up mechanism to focus on both local geometry details and global shape features of inputs. In addition, we design a scene-graph in the network to pay attention to the missing parts by the aid of neighboring vehicles. Experiments show that TraPCC achieve good performance for real-scan completion on KITTI and nuScenes traffic datasets even without any complete data in training. We also show a downstream application of 3D detection, which benefits from our completion approach. © 2022 IEEE.
会议录编者/会议主办者	CAS ; IEEE ; IEEE Circuits and Systems Society (CAS) ; IEEE Communications Society (ComSoc) ; IEEE Signal Processing Society
关键词	Computer vision Deep learning Large dataset 'current 3-d scans Bottom up Learning-based approach Partial geometry Point-clouds Real traffic Self occlusion Sensor resolution Traffic scene
会议名称	2022 IEEE International Conference on Multimedia and Expo, ICME 2022
会议地点	Taipei, Taiwan
会议日期	July 18, 2022 - July 22, 2022
URL	查看原文
收录类别	EI
语种	英语
出版者	IEEE Computer Society
EI入藏号	20223712732781
EI主题词	Vehicles
EISSN	1945-788X
EI分类号	461.4 Ergonomics and Human Factors Engineering ; 723.2 Data Processing and Image Processing ; 723.5 Computer Applications ; 741.2 Vision
原始文献类型	Conference article (CA)
来源库	IEEE
文献类型	会议论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/232000
专题	信息科学与技术学院_博士生信息科学与技术学院_硕士生信息科学与技术学院_PI研究组_马月昕
作者单位	1.ShanghaiTech University 2.Chinese University of Hong Kong
第一作者单位	上海科技大学
第一作者的第一单位	上海科技大学
推荐引用方式 GB/T 7714	Yiming Ren,Peishan Cong,Xinge Zhu,et al. Self-Supervised Point Cloud Completion on Real Traffic Scenes Via Scene-Concerned Bottom-Up Mechanism[C]//CAS, IEEE, IEEE Circuits and Systems Society (CAS), IEEE Communications Society (ComSoc), IEEE Signal Processing Society:IEEE Computer Society,2022.