Home

Awesome

P2B: Point-to-Box Network for 3D Object Tracking in Point Clouds

Introduction

This repository is released for P2B in our CVPR 2020 paper (oral). Here we include our P2B model (PyTorch) and code for data preparation, training and testing on KITTI tracking dataset.

Preliminary

    pip install -r requirements.txt
    python setup.py build_ext --inplace

Evaluation

Train a new P2B model on KITTI data:

python train_tracking.py --data_dir=<kitti data path>

Test a new P2B model on KITTI data:

python test_tracking.py --data_dir=<kitti data path>

Please refer to the code for setting of other optional arguments, including data split, training and testing parameters, etc.

Acknowledgements

Thank Giancola for his implementation of SC3D. Thank Erik Wijmans for his implementation of PointNet++ in PyTorch. Thank Charles R. Qi for his implementation of Votenet. They help and inspire this work.