GroupContrast

This repo is the official project repository of the paper GroupContrast: Semantic-aware Self-supervised Representation Learning for 3D Understanding and is mainly used for releasing schedules, updating instructions, sharing experiment records (contains model weight), and handling issues. The code will be updated in Pointcept.

[GroupContrast] - [ arXiv ] [ Code ]

Highlights

• Mar, 2023: We release the project repo for GroupContrast, if you have any questions related to our work, please feel free to open an issue.

Schedule

Our code release schedule is as follows:

• Release pre-training config and code of GroupContrast;
  • ScanNet
• Release semantic segmentation fine-tuning code and config;
  • ScanNet
  • ScanNet200
  • S3DIS
  • ScanNet data efficient
• Release instance segmentation fine-tuning code and config;
  • ScanNet
  • ScanNet200
  • S3DIS
• Release object detection fine-tuning code and config;
  • ScanNet
  • SUN-RGBD

Citation

If you find GroupContrast useful to your research, please cite our work as an acknowledgment.

@inproceedings{wang2023gc,
    title={GroupContrast: Semantic-aware Self-supervised Representation Learning for 3D Understanding}, 
    author={Wang, Chengyao and Jiang, Li and Wu, Xiaoyang and Tian, Zhuotao and Peng, Bohao and Zhao, Hengshuang and Jia, Jiaya},
    booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    year={2024}
}

@inproceedings{wu2024ppt,
    title={Towards Large-scale 3D Representation Learning with Multi-dataset Point Prompt Training},
    author={Wu, Xiaoyang and Tian, Zhuotao and Wen, Xin and Peng, Bohao and Liu, Xihui and Yu, Kaicheng and Zhao, Hengshuang},
    booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    year={2024}
}

@inproceedings{wu2023msc,
  title={Masked Scene Contrast: A Scalable Framework for Unsupervised 3D Representation Learning},
  author={Wu, Xiaoyang and Wen, Xin and Liu, Xihui and Zhao, Hengshuang},
  journal={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2023}
}

@misc{pointcept2023,
    title={Pointcept: A Codebase for Point Cloud Perception Research},
    author={Pointcept Contributors},
    howpublished={\url{https://github.com/Pointcept/Pointcept}},
    year={2023}
}