CenterPoint

3D Object Detection and Tracking using center points in the bird-eye view.

This forked repo were used for the lecture Advanced Deep Learning for Computer Vision. Our poster and contributions:

Reference papers \ Repos:

1. Center Point: https://arxiv.org/abs/2006.11275
2. BiFPN horizon Robotics: https://arxiv.org/abs/2006.15505
3. BiFPN GitHub Repo: https://github.com/ViswanathaReddyGajjala/EfficientDet-Pytorch
4. Metrics GitHub repo: https://github.com/rafaelpadilla/Object-Detection-Metrics

GCP VM Settings:

• Intel N1-highmem-4 (4vCPU, 26 GB Memory)
• Nvidia Tesla T4
• 500 GB SSD

Useful Links:

1. GCP GPU Zones: https://cloud.google.com/compute/docs/gpus/gpu-regions-zones

Original Repo README.md:

3D Object Detection and Tracking using center points in the bird-eye view.

Center-based 3D Object Detection and Tracking,
Tianwei Yin, Xingyi Zhou, Philipp Krähenbühl,
arXiv technical report (arXiv 2006.11275)

@article{yin2021center,
  title={Center-based 3D Object Detection and Tracking},
  author={Yin, Tianwei and Zhou, Xingyi and Kr{\"a}henb{\"u}hl, Philipp},
  journal={CVPR},
  year={2021},
}

This repo is an reimplementation of CenterPoint on the KITTI dataset. For nuScenes and Waymo, please refer to the original repo. Please refer to INSTALL.md for installation. We provide two configs, centerpoint.yaml for the vanilla centerpoint model and centerpoint_rcnn.yaml which combines centerpoint with PVRCNN. Pretrained models are coming soon.

Acknowledgement

Our code is based on OpenPCDet. Some util files are copied from mmdetection and mmdetection3d. Thanks OpenMMLab Development Team for their awesome codebases.