Our paper is accepted by IJCAI2024. Thanks for the OpenPCDet, this implementation of the DCDet is mainly based on the pcdet v0.6.
Abstract: Recently, significant progress has been made in the research of 3D object detection. However, most prior studies have focused on the utilization of center-based or anchor-based label assignment schemes. Alternative label assignment strategies remain unexplored in 3D object detection. We find that the center-based label assignment often fails to generate sufficient positive samples for training, while the anchor-based label assignment tends to encounter an imbalanced issue when handling objects of varying scales. To solve these issues, we introduce a dynamic cross label assignment (DCLA) scheme, which dynamically assigns positive samples for each object from a cross-shaped region, thus providing sufficient and balanced positive samples for training. Furthermore, to address the challenge of accurately regressing objects with varying scales, we put forth a rotation-weighted Intersection over Union (RWIoU) metric to replace the widely used L1 metric in regression loss. Extensive experiments demonstrate the generality and effectiveness of our DCLA and RWIoU-based regression loss.
- Linux (tested on Ubuntu 20.04)
- Python 3.6+
- PyTorch 1.1 or higher (tested on PyTorch 1.13)
- CUDA 9.0 or higher (tested on 11.6)
pip install -r requirement.txt
python setup.py develop- Prepare KITTI dataset and road planes
# Download KITTI and organize it into the following form:
├── data
│ ├── kitti
│ │ │── ImageSets
│ │ │── training
│ │ │ ├──calib & velodyne & label_2 & image_2 & (optional: planes)
│ │ │── testing
│ │ │ ├──calib & velodyne & image_2
# Generatedata infos:
python -m pcdet.datasets.kitti.kitti_dataset create_kitti_infos tools/cfgs/dataset_configs/kitti_dataset.yaml- Prepare Waymo dataset
# Download Waymo and organize it into the following form:
├── data
│ ├── waymo
│ │ │── ImageSets
│ │ │── raw_data
│ │ │ │── segment-xxxxxxxx.tfrecord
| | | |── ...
| | |── waymo_processed_data
│ │ │ │── segment-xxxxxxxx/
| | | |── ...
│ │ │── pcdet_gt_database_train_sampled_xx/
│ │ │── pcdet_waymo_dbinfos_train_sampled_xx.pkl
# Install tf 2.5.0
# Install the official waymo-open-dataset by running the following command:
pip3 install --upgrade pip
pip3 install waymo-open-dataset-tf-2-5-0 --user
# Extract point cloud data from tfrecord and generate data infos:
python -m pcdet.datasets.waymo.waymo_dataset --func create_waymo_infos --cfg_file tools/cfgs/dataset_configs/waymo_dataset.yaml- Train with a single GPU
python tools/train.py --cfg_file ${CONFIG_FILE}
# e.g.,
python tools/train.py --cfg_file tools/cfgs/waymo_models/dcdet.yaml- Train with multiple GPUs or multiple machines
bash tools/scripts/dist_train.sh ${NUM_GPUS} --cfg_file ${CONFIG_FILE}
# or
bash tools/scripts/slurm_train.sh ${PARTITION} ${JOB_NAME} ${NUM_GPUS} --cfg_file ${CONFIG_FILE}
# e.g.,
bash tools/scripts/dist_train.sh 8 --cfg_file tools/cfgs/waymo_models/dcdet.yaml- Test with a pretrained model:
python tools/test.py --cfg_file ${CONFIG_FILE} --ckpt ${CKPT}
# e.g.,
python tools/test.py --cfg_file tools/cfgs/waymo_models/dcdet.yaml --ckpt {path}Please cite our paper if you find our work useful for your research:
@inproceedings{ijcai2024p124,
title = {DCDet: Dynamic Cross-based 3D Object Detector},
author = {Liu, Shuai and Li, Boyang and Fang, Zhiyu and Huang, Kai},
booktitle = {Proceedings of the Thirty-Third International Joint Conference on
Artificial Intelligence, {IJCAI-24}},
publisher = {International Joint Conferences on Artificial Intelligence Organization},
editor = {Kate Larson},
pages = {1119--1127},
year = {2024},
month = {8},
note = {Main Track},
doi = {10.24963/ijcai.2024/124},
url = {https://doi.org/10.24963/ijcai.2024/124},
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent