xmuqimingxia / CoIn

CoIn: Contrastive Instance Feature Mining for Outdoor 3D Object Detection with Very Limited Annotations(ICCV2023)

CoIn: Contrastive Instance Feature Mining for Outdoor 3D Object Detection with Very Limited Annotations(ICCV2023)

This is a official code release of CoIn (Contrastive Instance Feature Mining for Outdoor 3D Object Detection with Very Limited Annotations). This code is mainly based on OpenPCDet.

Detection Framework

Getting Started

Prepare dataset

We provide a kitti_ Info_train.pkl for running code. https://drive.google.com/file/d/1342-YP-w8xbC2PDGcaoSASaJkjHKN89u/view?usp=sharing

Training & Testing

Train a model

You could optionally add extra command line parameters --batch_size ${BATCH_SIZE} and --epochs ${EPOCHS} to specify your preferred parameters.

• Train with multiple GPUs or multiple machines

sh scripts/dist_train.sh ${NUM_GPUS} tools/cfgs/kitti_models/CoIn.yaml

• Train with a single GPU:

python train.py tools/cfgs/kitti_models/CoIn.yaml

Test a model

• To test with multiple GPUs:

sh scripts/dist_test.sh ${NUM_GPUS} \
    tools/cfgs/kitti_models/CoIn.yaml --batch_size ${BATCH_SIZE}

• To test all the saved checkpoints of a specific training setting and draw the performance curve on the Tensorboard, add the --eval_all argument:

python test.py tools/cfgs/kitti_models/CoIn.yaml --batch_size ${BATCH_SIZE} --eval_all

Acknowledgement

OpenPCDET

SIOD

Citation

If you find this project useful in your research, please consider cite:

@inproceedings{CoIn2023,
    title={CoIn: Contrastive Instance Feature Mining for Outdoor 3D Object Detection with Very Limited Annotations},
    author={Xia, Qiming and Deng, Jinhao and Wen, Chenglu and Wu, Hai and Shi, Shaoshuai and Li, Xin and Wang, Cheng},
    booktitle = {ICCV},
    year={2023}
}