mtubpeng1 / cmignet

Cross-Modal Information-Guided Network using Contrastive Learning for Point Cloud Registration (RAL 2023)

Cross-Modal Information-Guided Network using Contrastive Learning for Point Cloud Registration (RAL 2023)

PyTorch implementation of the paper: Cross-Modal Information-Guided Network using Contrastive Learning for Point Cloud Registration.

Environment

Our model is trained with the following environment:

• Ubuntu 20.04
• Python 3.8
• PyTorch 1.8.1 with torchvision 0.9.1 (Cuda 11.1) Other required packages can be found in requirements.txt.

Dataset Preparation

The cross-modal ModelNet40 dataset can be downloaded from Google Drive. You can download and unzip it to the data folder.

The pre-trained models can be downloaded from Google Drive.

Usage

You can see a list of options that can be used to control hyperparameters of the model and experiment settings at the end of main.py. The comments in the file should be enough to understand them.

To train a model:

python main.py 

To test a model:

python test.py --model_path <path_to_model>

Citation

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

@article{xie2023cross,
  author={Xie, Yifan and Zhu, Jihua and Li, Shiqi and Shi, Pengcheng},
  journal={IEEE Robotics and Automation Letters}, 
  title={Cross-Modal Information-Guided Network Using Contrastive Learning for Point Cloud Registration}, 
  year={2024},
  volume={9},
  number={1},
  pages={103-110},
  doi={10.1109/LRA.2023.3331625}
}