https://github.com/gondar-software/pointnetlk
https://github.com/gondar-software/pointnetlk
Last synced: about 1 year ago
JSON representation
- Host: GitHub
- URL: https://github.com/gondar-software/pointnetlk
- Owner: gondar-software
- License: mit
- Created: 2025-03-21T19:40:34.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-03-21T22:31:32.000Z (about 1 year ago)
- Last Synced: 2025-03-25T21:35:31.210Z (about 1 year ago)
- Language: Python
- Size: 66.4 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# PointNetLK: Point Cloud Registration using PointNet
### [Video](https://youtu.be/J2ClR5OZuLc)
Source Code Author:
Yasuhiro Aoki
### Requires:
* PyTorch 0.4.0 (perhaps, 0.4.1 (the latest) will be OK.) and torchvision
* NumPy
* SciPy
* MatPlotLib
* ModelNet40
### Main files for experiments:
* train_classifier.py: train PointNet classifier (used for transfer learning)
* train_pointlk.py: train PointNet-LK
* generate_rotation.py: generate 6-dim perturbations (rotation and translation) (for testing)
* test_pointlk.py: test PointNet-LK
* test_icp.py: test ICP
* result_stat.py: compute mean errors of above tests
### Examples (Bash shell scripts):
* ex1_train.sh: train PointNet classifier and transfer to PointNet-LK.
* ex1_genrot.sh: generate perturbations for testing
* ex1_test_pointlk.sh: test PointNet-LK
* ex1_test_icp.sh: test ICP
* ex1_result_stat.sh: compute mean errors of above tests
### Citation
```
@InProceedings{yaoki2019pointnetlk,
author = {Aoki, Yasuhiro and Goforth, Hunter and Arun Srivatsan, Rangaprasad and Lucey, Simon},
title = {PointNetLK: Robust & Efficient Point Cloud Registration Using PointNet},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2019}
}
```