PartNet: A Large-scale Benchmark for Fine-grained and Hierarchical Part-level 3D Object Understanding

Figure 1. The PartNet Annotation System Overview.

Annotation System

This repo contains the web-based part segmentation annotation interface for PartNet.

Run

1. docker-compose build
2. docker-compose up
3. If up fails once then run docker-compose down
4. Rerun docker-compose up
5. If you would like to use nodemon then replace node ./bin/www with nodemon ./bin/www in server/package.json and follow steps 1 and 2 again.

Paper and Dataset

PartNet is accepted to CVPR 2019. See you at Long Beach, CA.

Our team: Kaichun Mo, Shilin Zhu, Angel X. Chang, Li Yi, Subarna Tripathi, Leonidas J. Guibas and Hao Su from Stanford, UCSD, SFU and Intel AI Lab.

Arxiv Version: https://arxiv.org/abs/1812.02713

Project Page: https://cs.stanford.edu/~kaichun/partnet/

Video: https://youtu.be/7pEuoxmb-MI

Please refer to this repo for the PartNet dataset utilities and this repo for the segmentation experiments (Section 5) in the paper.

Citations

@InProceedings{Mo_2019_CVPR,
    author = {Mo, Kaichun and Zhu, Shilin and Chang, Angel X. and Yi, Li and Tripathi, Subarna and Guibas, Leonidas J. and Su, Hao},
    title = {{PartNet}: A Large-Scale Benchmark for Fine-Grained and Hierarchical Part-Level {3D} Object Understanding},
    booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
    month = {June},
    year = {2019}
}

License

MIT Licence

Updates

• [April 18, 2019] PartNet Annotation System v1.0 release.