Code and Dataset for Convolutional Occupancy Models for Dense Packing of Complex, Novel Objects

This repo contains the dataset and model proposed in our IROS 2023 paper:

"Convolutional Occupancy Models for Dense Packing of Complex, Novel Objects".
Nikhil Mishra, Pieter Abbeel, Xi Chen, and Maximilan Sieb.
In the proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2023.

The dataset, COB-3D-v2, can be downloaded from our project page.

The shape-completion model we proposed, F-CON, is implemented in fcon_model.py.

The Jupyter notebook example.ipynb shows how to load the data and do inference with a pretrained F-CON model we have provided.

Citations

If you use COB-3D-v2 or F-CON in your work, please cite:

@inproceedings{
    mishra2023convolutional,
    title={Convolutional Occupancy Models for Dense Packing of Complex, Novel Objects},
    author={Nikhil Mishra and Pieter Abbeel and Xi Chen and Maximilian Sieb},
    year={2023},
    booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
}

License

Shield:

This work, including the paper, code, weights, and dataset, is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.