V2PNet: Voxel-to-Point Feature Propagation and Fusion that Improves Feature Representation for Point Cloud Registration

Point-based and voxel-based methods can learn the local features of point clouds. However, although point-based methods are geometrically precise, the discrete nature of point clouds negatively affects feature learning performance. Moreover, although voxel-based methods can exploit the learning powe...

Full description

Saved in:
Bibliographic Details
Published in:IEEE journal of selected topics in applied earth observations and remote sensing 2023-01, Vol.16, p.1-12
Main Authors: Hu, Han, Hou, Yongkuo, Ding, Yulin, Pan, Guoqiang, Chen, Min, Ge, Xuming
Format: Article
Language:English
Subjects:
3-D feature point
3-D metric learning
3D feature point
3D metric learning
Artificial neural networks
Coders
Convolution
Datasets
Feature extraction
Localization
Methods
Neural networks
Performance measurement
Point cloud compression
point cloud registration
Propagation
Representation learning
sparse convolution
Task analysis
Three-dimensional displays
Work platforms
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Point-based and voxel-based methods can learn the local features of point clouds. However, although point-based methods are geometrically precise, the discrete nature of point clouds negatively affects feature learning performance. Moreover, although voxel-based methods can exploit the learning power of convolutional neural networks, their resolution and detail extraction may be inadequate. Therefore, in this study, point-based and voxel-based methods are combined to enhance localization precision and matching distinctiveness. The core procedure is embodied in V2PNet, an innovative fused neural network that we design to perform voxel-to-pixel propagation and fusion, which seamlessly integrates the two encoder-decoder branches. Experiments are conducted on indoor and outdoor benchmark datasets with different platforms and sensors, i.e., the 3DMatch and KITTI datasets, with the registration recall of 89.4% and the success rate of 99.86%, respectively. Qualita-tive and quantitative evaluations demonstrate that V2PNet has shown improvements in semantic awareness, geometric structure discernment, and other performance metrics. Codes are made public at https://github.com/houyongkuo/V2PNet .
ISSN:1939-1404
2151-1535
DOI:10.1109/JSTARS.2023.3278830