中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
  5. 先进控制与自动化团队
3D-RVP: A method for 3D object reconstruction from a single depth view using voxel and point

文献类型：期刊论文

作者Zhao, Meihua3,4; Xiong, Gang5,6; Zhou, MengChu1,7; Shen, Zhen2,3; Wang, Fei-Yue3
刊名NEUROCOMPUTING
出版日期2021-03-21
卷号430页码:94-103
关键词3D object reconstruction Encoder-decoder network Machine learning Point prediction network
ISSN号0925-2312
DOI10.1016/j.neucom.2020.10.097
通讯作者Shen, Zhen(zhen.shen@ia.ac.cn)
英文摘要Three-dimensional object reconstruction technology has a wide range of applications such as augment reality, virtual reality, industrial manufacturing and intelligent robotics. Although deep learning-based 3D object reconstruction technology has developed rapidly in recent years, there remain important problems to be solved. One of them is that the resolution of reconstructed 3D models is hard to improve because of the limitation of memory and computational efficiency when deployed on resource-limited devices. In this paper, we propose 3D-RVP to reconstruct a complete and accurate 3D geometry from a single depth view, where R, V and P represent Reconstruction, Voxel and Point, respectively. It is a novel two-stage method that combines a 3D encoder-decoder network with a point prediction network. In the first stage, we propose a 3D encoder-decoder network with residual learning to output coarse prediction results. In the second stage, we propose an iterative subdivision algorithm to predict the labels of adaptively selected points. The proposed method can output high-resolution 3D models by increasing a small number of parameters. Experiments are conducted on widely used benchmarks of a ShapeNet dataset in which four categories of models are selected to test the performance of neural networks. Experimental results show that our proposed method outperforms the state-of-the-arts, and achieves about 2:7% improvement in terms of the intersection-over-union metric. (c) 2020 Elsevier B.V. All rights reserved.
资助项目National Natural Science Foundation of China[61773382] ; National Natural Science Foundation of China[U1909218] ; National Natural Science Foundation of China[U1909204] ; National Natural Science Foundation of China[61773381] ; National Natural Science Foundation of China[U1811463] ; National Natural Science Foundation of China[61872365] ; National Natural Science Foundation of China[61806198] ; CAS Key Technology Talent Program ; Chinese Guangdong's ST Project[2019B1515120030]
WOS研究方向Computer Science
语种英语
WOS记录号WOS:000617365300009
出版者ELSEVIER
资助机构National Natural Science Foundation of China ; CAS Key Technology Talent Program ; Chinese Guangdong's ST Project
源URL[http://ir.ia.ac.cn/handle/173211/43216]  
专题自动化研究所_复杂系统管理与控制国家重点实验室_先进控制与自动化团队
通讯作者Shen, Zhen
作者单位1.Macau Univ Sci & Technol, Inst Syst Engn, Macau, Peoples R China
2.Qingdao Acad Intelligent Ind, Intelligent Mfg Ctr, Qingdao 266109, Peoples R China
3.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing 100190, Peoples R China
4.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China
5.Chinese Acad Sci, Beijing Engn Res Ctr Intelligent Syst & Technol, Inst Automat, Beijing 100190, Peoples R China
6.Chinese Acad Sci, Guangdong Engn Res Ctr 3D Printing & Intelligent, Cloud Comp Ctr, Dongguan 523808, Peoples R China
7.New Jersey Inst Technol, Helen & John C Hartmann Dept Elect & Comp Engn, Newark, NJ 07102 USA
推荐引用方式
GB/T 7714		 Zhao, Meihua,Xiong, Gang,Zhou, MengChu,et al. 3D-RVP: A method for 3D object reconstruction from a single depth view using voxel and point[J]. NEUROCOMPUTING,2021,430:94-103.
APA Zhao, Meihua,Xiong, Gang,Zhou, MengChu,Shen, Zhen,&Wang, Fei-Yue.(2021).3D-RVP: A method for 3D object reconstruction from a single depth view using voxel and point.NEUROCOMPUTING,430,94-103.
MLA Zhao, Meihua,et al."3D-RVP: A method for 3D object reconstruction from a single depth view using voxel and point".NEUROCOMPUTING 430(2021):94-103.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。