Skip to content

Code for "BiCo-Net: Regress Globally, Match Locally for Robust 6D Pose Estimation"

License

Notifications You must be signed in to change notification settings

Gorilla-Lab-SCUT/BiCo-Net

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

BiCo-Net

Code for "BiCo-Net: Regress Globally, Match Locally for Robust 6D Pose Estimation" [Arxiv][Paper]

image

Requirements

This code has been tested with

  • Open3D==0.9.0.0
  • Python==3.7.12
  • OpenCV==4.1
  • Pytorch==1.6.0
  • CUDA==10.1

Downloads

  • YCB-Video dataset [link]
  • Preprocessed LineMOD dataset provided by DenseFusion [link]
  • Pretrained models of BiCo-Net [link]
  • Pred mask of PVN3D on YCB-Video dataset [link]
  • Pred mask of HybridPose on LineMOD Occlusion dataset [link]

Training

Command for training BiCo-Net:

python train.py --dataset ycbv --dataset_root path_to_ycbv_dataset
python train.py --dataset linemod --dataset_root path_to_lm_dataset

For lmo dataset, download the VOC2012 dataset and run:

python train.py --dataset lmo --dataset_root path_to_lm_dataset --bg_img path_to_voc2012_dataset

Evaluation

Evaluate the results of BiCo-Net reported in the paper:

python eval_ycbv.py --dataset_root path_to_ycbv_dataset --pred_mask path_to_pvn3d_pred_mask
python eval_lm.py --dataset_root path_to_lm_dataset
python eval_lmo.py --dataset_root path_to_lm_dataset --pred_mask path_to_hybridpose_pred_mask

Acknowledgements

Our implementation leverages the code from DenseFusion.

License

Our code is released under MIT License (see LICENSE file for details).

Citation

If you find our work useful in your research, please consider citing:

 @inproceedings{Xu2022BiCoNetRG,
   title={BiCo-Net: Regress Globally, Match Locally for Robust 6D Pose Estimation},
   author={Zelin Xu and Yichen Zhang and Ke Chen and Kui Jia},
   booktitle={IJCAI},
   year={2022}
 }

About

Code for "BiCo-Net: Regress Globally, Match Locally for Robust 6D Pose Estimation"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published