Uncertainty-aware Self-supervised 3D Data Association

This is the official repo of "Uncertainty-aware Self-supervised 3D Data Association", IROS 2020

Project page: https://jianrenw.github.io/Self-Supervised-3D-Data-Association/
Paper: https://arxiv.org/pdf/2008.08173.pdf

Data Preprocessing:

First generate sudo label for self-supervised embedding training:

bash ./tracking/main.sh
python ./embedding/dataset/nuscenes_preprocessing.py

Self-supervised embedding training:

bash ./embedding/main.sh

Combining appearance embedding with motion priors for MOT:

python ./combination/generate_data.py
python ./combination/logistic_regression.py
bash ./tracking/main.sh

Acknowlegement

This project is not possible without multiple great opensourced codebases. We list some notable examples below.

• AB3DMOT
• pointnet

SS3DA is deeply influenced by the following projects. Please consider citing the relevant papers.

@article{Weng2020_AB3DMOT, 
    author = {Weng, Xinshuo and Wang, Jianren and Held, David and Kitani, Kris}, 
    journal = {IROS}, 
    title = {{3D Multi-Object Tracking: A Baseline and New Evaluation Metrics}}, 
    year = {2020} 
}

@inproceedings{jianren20s3da,
    Author = {Wang, Jianren and Ancha, Siddharth and Chen, Yi-Ting and Held, David},
    Title = {Self-supervised 3D Data Association},
    Booktitle = {IROS},
    Year = {2020}
}