Hallucinated Hollow-3D R-CNN

This is the official implementation of From Multi-View to Hollow-3D: Hallucinated Hollow-3D R-CNN for 3D Object Detection, built on OpenPCDet. This paper has been accepted by IEEE TCSVT.

@article{deng2021hh3d,
  title={From Multi-View to Hollow-3D: Hallucinated Hollow-3D R-CNN for 3D Object Detection},
  author={Deng, Jiajun and Zhou, Wengang and Zhang, Yanyong and Li, Houqiang},
  journal={arXiv:2107.14391},
  year={2021}
}

Installation

1. Prepare for the running environment.

   You can either use the docker image we provide, or follow the installation steps in OpenPCDet.

   docker pull djiajun1206/pcdet:pytorch1.6
   

2. Prepare for the data.

   Please download the official KITTI 3D object detection dataset and organize the downloaded files as follows (the road planes could be downloaded from [road plane], which are optional for data augmentation in the training):

   Voxel-R-CNN
   ├── data
   │   ├── kitti
   │   │   │── ImageSets
   │   │   │── training
   │   │   │   ├──calib & velodyne & label_2 & image_2 & (optional: planes)
   │   │   │── testing
   │   │   │   ├──calib & velodyne & image_2
   ├── pcdet
   ├── tools
   

   Generate the data infos by running the following command:

   python -m pcdet.datasets.kitti.kitti_dataset create_kitti_infos tools/cfgs/dataset_configs/kitti_dataset.yaml
   

3. Setup.

   python setup.py develop
   

Getting Started

0. Downloading the model.

   The model reported in the manuscript can be download here.

1. Training.

   The configuration file is in tools/cfgs/kitti_models/hh3d_rcnn_car.yaml, and the training scripts is in tools/scripts.

   cd tools
   sh scripts/train_hh3d_rcnn.sh
   

2. Evaluation.

   The configuration file is in tools/cfgs/voxelrcnn, and the training scripts is in tools/scripts.

   cd tools
   sh scripts/eval_hh3d_rcnn.sh
   

Acknowledge

Thanks to the strong and flexible OpenPCDet codebase.