README.md

Point Cloud Mamba: Point Cloud Learning via State Space Model

Tao Zhang, Haobo Yuan, Lu Qi, Jiangning Zhang, Qianyu Zhou, Shunping Ji, Shuicheng Yan, Xiangtai Li

News

• PCM is finally accepted by AAAI-2025.
• All codes and weights are available at SkyworkAI/PointCloudMamba.

Features

• PCM introduces Mamba to point cloud analysis.
• PCM can perform global modeling while maintaining linear computational complexity.
• PCM outperforms PointNeXt on the ScanObjectNN, ModelNet40, and ShapeNetPart datasets.

Demos

ShapeNetPart

Performance

3-D Point Cloud Classification

3-D Point Cloud Segmentation

@article{zhang2024point,
  title={Point Cloud Mamba: Point Cloud Learning via State Space Model},
  author={Zhang, Tao and Yuan, Haobo and Qi, Lu and Zhanng, Jiangning and Zhou, Qianyu and Ji, Shunping and Yan, Shuicheng and Li, Xiangtai},
  journal={AAAI},
  year={2025}
}

License

MIT License