Skip to content
This repository was archived by the owner on Jun 4, 2023. It is now read-only.
/ d3d Public archive

Devkit for 3D -- Some utils for 3D object detection based on Numpy and Pytorch

License

Notifications You must be signed in to change notification settings

cmpute/d3d

Folders and files

NameName
Last commit message
Last commit date
Feb 5, 2021
May 14, 2020
Feb 7, 2021
Feb 17, 2022
May 19, 2021
Feb 14, 2021
Sep 9, 2021
Aug 20, 2021
May 14, 2020
Aug 20, 2021
May 16, 2021
Feb 3, 2020
Jan 26, 2021
Mar 1, 2021
Aug 11, 2020
Jan 26, 2021
May 19, 2021
May 19, 2021

Repository files navigation

D3D

Devkit for 3D: Some utils for 3D object detection and tracking based on Numpy and Pytorch

PyPI Version Docs Status


Please consider siting my work if you find this library useful in your research :)
@article{zhong2020uncertainty,
  title={Uncertainty-Aware Voxel based 3D Object Detection and Tracking with von-Mises Loss},
  author={Zhong, Yuanxin and Zhu, Minghan and Peng, Huei},
  journal={arXiv preprint arXiv:2011.02553},
  year={2020}
}

Features

  • Unified data representation
  • Support loading KITTI, Waymo, Nuscenes dataset
  • Rotated 2D IoU, NMS with clear CUDA implementations
  • Point Cloud Voxelization
  • Visualization
  • Benchmarking

Tips

  • Current polygon intersecting algorithm is not very stable, so try to convert the input to double precision if you met error with the iou functions