Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
Zhengping Che 6f187a3e7d | 1 year ago | |
---|---|---|
configs | 1 year ago | |
data | 1 year ago | |
fig | 1 year ago | |
mmdet3d | 1 year ago | |
requirements | 1 year ago | |
tools | 1 year ago | |
.gitignore | 1 year ago | |
LICENSE | 1 year ago | |
README.md | 1 year ago | |
requirements.txt | 1 year ago | |
setup.cfg | 1 year ago | |
setup.py | 1 year ago |
This is a release of the code of our paper Label-Guided Auxiliary Training Improves 3D Object Detector, ECCV 2022.
Authors: Yaomin Huang, Xinmei Liu, Yichen Zhu, Zhiyuan Xu, Chaomin Shen *, Zhengping Che, Guixu Zhang,
Yaxin Peng, Feifei Feng, and Jian Tang * (*corresponding author)
In this repository, we reimplement LG3D based on mmdetection3d for easier usage.
In this
paper, we propose a Label-Guided auxiliary training method for 3D object detection (LG3D in short), which serves as an auxiliary network to enhance the feature learning of existing 3D object detectors.
If you find our work useful in your research, please consider citing:
@inproceedings{huang2022label,
title={Label-guided auxiliary training improves 3d object detector},
author={Huang, Yaomin and Liu, Xinmei and Zhu, Yichen and Xu, Zhiyuan and Shen, Chaomin and Che, Zhengping and Zhang, Guixu and Peng, Yaxin and Feng, Feifei and Tang, Jian},
booktitle={Computer Vision--ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23--27, 2022, Proceedings, Part IX},
pages={684--700},
year={2022},
organization={Springer}
}
This repo is built based on mmdetection3d (V1.0.0), please follow the getting_started.md for installation.
The code is tested under the following environment:
Please follow the instruction here to prepare ScanNet Data.
Please follow the instruction here to prepare SUN RGB-D Data.
For ScanNet V2, please run to training LG3D:
./tools/dist_train.sh configs/votenet/votenet_lg3d_scannet.py 4 --work-dir ./log/scannet/lg3d/
After training, please run :
python .tools/convert_fully2single.py
convert fully LG3D model to VoteNet.
Test VoteNet on ScanNet and evaluate the mAP.
python tools/test.py configs/votenet/votenet_8x8_scannet-3d-18class.py log/pre_trained/votenet_scannet_final.pth --eval mAP --eval-options 'out_dir=./log/scannet/show_results'
Test VoteNet on ScanNet and save the points and prediction visualization results.
python tools/test.py configs/votenet/votenet_8x8_scannet-3d-18class.py log/pre_trained/votenet_scannet_final.pth --show --show-dir ./log/scannet/show_results
Our code is heavily based on MMDetection3D. Thanks mmdetection3d Development Team for their awesome codebase.
The following modifications have been made to the original code:
configs/votenet/votenet_lg3d_scannet.py
mmdet3d/model/detector/lg3d_votenet.py
mmdet3d/models/backbone/attention_utils.py
Please ensure compliance with the licensing terms of the original project, i.e., the Apache License.
No Description
Python Text MATLAB other
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》