3D Point Cloud Restoration via Deep Learning: A Comprehensive Survey

被引：0
作者：
Liu C. [1 ]
Wei M. [2 ]
Guo Y. [3 ]
机构：
[1] Institute of EduInfo Science & Engineering, Nanjing Normal University, Nanjing
[2] College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing
[3] State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing
来源：
Wei, Mingqiang (mqwei@nuaa.edu.cn) | 1936年 / Institute of Computing Technology卷 / 33期
关键词：
3D point cloud; Deep learning; Feature coding; Network module; Point cloud restoration;
D O I：
10.3724/SP.J.1089.2021.18817
中图分类号：
学科分类号：
摘要：
3D point cloud is one of the most commonly used 3D scene/object representation methods. According to the dif-ferent emphases of point cloud restoration, 3D point cloud restoration technologies based on deep learning are divided into three classes: dense reconstruction, complete reconstruction and denoising reconstruction. Typical restoration models and key techniques, such as feature coding, feature extension, and loss function design, are analyzed. Commonly used network modules, point cloud data sets, and evaluation criteria are summarized. Finally, the relationship between the three kinds of point cloud restoration technologies is discussed, and the challenges and future development trends of point cloud restoration technologies are explored from five aspects: rotation invariant feature extraction, detail information repair, topological relationship preservation, geometric algorithm application, and multimodal data fusion. © 2021, Beijing China Science Journal Publishing Co. Ltd. All right reserved.
引用
收藏
页码：1936 / 1952
页数：16
相关论文
共 77 条
[71]  
Alexa M, Behr J, Cohen-Or D, Et al., Computing and rendering point set surfaces, IEEE Transactions on Visualization and Computer Graphics, 9, 1, pp. 3-15, (2003)
[72]  
Sarkar K, Varanasi K, Stricker D., Learning quadrangulated patches for 3D shape parameterization and completion, Proceedings of International Conference on 3D Vision, pp. 383-392, (2017)
[73]  
Lu X Q, Wu S H, Chen H H, Et al., GPF: GMM-inspired feature-preserving point set filtering, IEEE Transactions on Visualization and Computer Graphics, 24, 8, pp. 2315-2326, (2018)
[74]  
Qian Y, Hou J H, Kwong S, Et al., PUGeo-Net: a geometry-centric network for 3D point cloud upsampling
[75]  
He Jun, Zhang Caiqing, Li Xiaozhen, Et al., Survey of research on multimodal fusion technology for deep learning, Computer Engineering, 46, 5, pp. 1-11, (2020)
[76]  
Zhang Y D, Funkhouser T., Deep depth completion of a single RGB-D image, Proceedings of the 31st IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 175-185, (2018)
[77]  
Qiu J X, Cui Z P, Zhang Y D, Et al., DeepLiDAR: deep surface normal guided depth prediction for outdoor scene from sparse lidar data and single color image, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3313-3322, (2019)
12345678