jevonmao / dcp

Deep Closest Point

Prerequisites

PyTorch>=1.0: https://pytorch.org

scipy>=1.2

numpy

h5py

tqdm

TensorboardX: https://github.com/lanpa/tensorboardX

Training

DCP-v1

python main.py --exp_name=dcp_v1 --model=dcp --emb_nn=dgcnn --pointer=identity --head=svd

DCP-v2

python main.py --exp_name=dcp_v2 --model=dcp --emb_nn=dgcnn --pointer=transformer --head=svd

Testing

DCP-v1

python main.py --exp_name=dcp_v1 --model=dcp --emb_nn=dgcnn --pointer=identity --head=svd --eval

or

python main.py --exp_name=dcp_v1 --model=dcp --emb_nn=dgcnn --pointer=identity --head=svd --eval --model_path=xx/yy

DCP-v2

python main.py --exp_name=dcp_v2 --model=dcp --emb_nn=dgcnn --pointer=transformer --head=svd --eval

or

python main.py --exp_name=dcp_v2 --model=dcp --emb_nn=dgcnn --pointer=transformer --head=svd --eval --model_path=xx/yy

where xx/yy is the pretrained model

Citation

Please cite this paper if you want to use it in your work,

@InProceedings{Wang_2019_ICCV,
  title={Deep Closest Point: Learning Representations for Point Cloud Registration},
  author={Wang, Yue and Solomon, Justin M.},
  booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
  month = {October},
  year={2019}
}

License

MIT License