GeoFormer: Learning Point Cloud Completion with Tri-Plane Integrated Transformer

doi:10.1145/3664647.3680842

	GeoFormer: Learning Point Cloud Completion with Tri-Plane Integrated Transformer
	Yu, Jinpeng1,2 ; Huang, Binbin2 ; Zhang, Yuxuan 3; Li, Huaxia 1; Tang, Xu 1; Gao, Shenghua 4,5
	2024-10-28
会议录名称	MM 2024 - PROCEEDINGS OF THE 32ND ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA
页码	8952-8961
DOI	10.1145/3664647.3680842
摘要	Point cloud completion aims to recover accurate global geometry and preserve fine-grained local details from partial point clouds. Conventional methods typically predict unseen points directly from 3D point cloud coordinates or use self-projected multi-view depth maps to ease this task. However, these gray-scale depth maps cannot reach multi-view consistency, consequently restricting the performance. In this paper, we introduce a GeoFormer that simultaneously enhances the global geometric structure of the points and improves the local details. Specifically, we design a CCM Feature Enhanced Point Generator to integrate image features from multi-view consistent canonical coordinate maps (CCMs) and align them with pure point features, thereby enhancing the global geometry feature. Additionally, we employ the Multi-scale Geometry-aware Upsampler module to progressively enhance local details. This is achieved through cross attention between the multi-scale features extracted from the partial input and the features derived from previously estimated points. Extensive experiments on the PCN, ShapeNet-55/34, and KITTI benchmarks demonstrate that our GeoFormer outperforms recent methods, achieving the state-of-the-art performance. Our code is available at https://github.com/Jinpeng-Yu/GeoFormer. © 2024 Owner/Author.
会议录编者/会议主办者	ACM SIGMM
关键词	Canonical coordinate map Canonical coordinates Coordinate maps Depthmap Multi-scale geometry-aware Multi-scales Multi-view consistent Multi-views Point cloud completion Point-clouds
会议名称	32nd ACM International Conference on Multimedia, MM 2024
会议地点	Melbourne, VIC, Australia
会议日期	October 28, 2024 - November 1, 2024
URL	查看原文
收录类别	EI
语种	英语
出版者	Association for Computing Machinery, Inc
EI入藏号	20244817416924
原始文献类型	Conference article (CA)
文献类型	会议论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/455187
专题	信息科学与技术学院_硕士生信息科学与技术学院_博士生
通讯作者	Gao, Shenghua
作者单位	1.Xiaohongshu Inc., Shanghai, China; 2.ShanghaiTech University, Shanghai, China; 3.Shanghai Jiao Tong University, Shanghai, China; 4.The University of Hong Kong, Hong Kong, Hong Kong; 5.HKU Shanghai Advanced Computing and Intelligent Technology Research Institute, China
第一作者单位	上海科技大学
推荐引用方式 GB/T 7714	Yu, Jinpeng,Huang, Binbin,Zhang, Yuxuan,et al. GeoFormer: Learning Point Cloud Completion with Tri-Plane Integrated Transformer[C]//ACM SIGMM:Association for Computing Machinery, Inc,2024:8952-8961.