Awesome
Rotation-invariant-deep-pointcloud-analysis
Code for the ICCV 2021 paper: A-closer-look-at-rotation-invariant-deep-pointcloud-analysis.
Authors: Feiran Li, Kent Fujiwara, Fumio Okura, and Yasuyuki Matsushita
1. Environment
The provide codes have been tested with Pytorch-1.6.0 on a Tesla-V100.
1. Run the code
- Download the PCA-processed datasets (ModelNet40, ShapeNet-PartSeg, and ScanObjectNN) and unzip them to the
dataset
folder. - Note that the
ScanObjectNN
dataset is originally provided here. Please pay attention to citation. - Run respective
*_test.ipynb
to test the pretrained model and*_train.ipynb
to train from scratch. - If you want to generate the 24 ambiguities of your own dataset, please see the
generate_24_pca_poses.py
script.
3. Contact
Please feel free to raise an issue or email to li.feiran@ist.osaka-u.ac.jp if you have any question regarding the paper or any suggestions for further improvements.
4. Citation
If you find this code helpful, thanks for citing our work as
@inproceedings{li2021rotinv,
title = {A Closer Look at Rotation-invariant Deep Point Cloud Analysis},
author = {Feiran Li and Kent Fujiwara and Fumio Okura and Yasuyuki Matsushita},
booktitle = {IEEE/CVF International Conference on Computer Vision (ICCV)},
year = {2021}
}