Uncertainty-Instructed Structure Injection for Generalizable HD Map Construction

Xiaolu Liu*, Ruizi Yang*, Song Wang, Wentong Li, Junbo Chen†, Jianke Zhu†

[Paper] (arXiv). CVPR2025

Video Demo

Introduction

We propose UIGenMap, an uncertainty-instructed structure injection approach for generalizable HD map vectorization, which concerns the uncertainty resampling in statistical distribution and employs explicit instance features to reduce excessive reliance on training data.

Getting Started

1. Environment

Step 1. Create conda environment and activate it.

conda create --name uigenmapnet python=3.8 -y
conda activate uigenmapnet

Step 2. Install PyTorch.

pip install torch==1.9.0+cu111 torchvision==0.10.0+cu111 torchaudio==0.9.0 -f https://download.pytorch.org/whl/torch_stable.html

Step 3. Install MMCV series.

# Install mmcv-series
pip install mmcv-full==1.6.0
pip install mmdet==2.28.2
pip install mmsegmentation==0.30.0
git clone https://github.com/open-mmlab/mmdetection3d.git
cd mmdetection3d
git checkout v1.0.0rc6 
pip install -e .

Step 4. Install other requirements.

pip install -r requirements.txt

2. nuScenes Dataset Preparation

Step 1. Download NuScenes dataset to ./data/nuscenes.

Step 2. Generate annotation files for NuScenes dataset.

python tools/nuscenes_converter.py --data-root ./data/nuscenes --newsplit

3. Training and Validating

To train a model with 2 GPUs:

bash tools/dist_train.sh ${CONFIG} 2

To validate a model with 2 GPUs:

bash tools/dist_test.sh ${CONFIG} ${CEHCKPOINT} 2 --eval

Pretrained pv_det.pth.

Result checkpoint and training log.

Acknowledgements

UIGenMap is based on mmdetection3d. It is also greatly inspired by the following outstanding contributions to the open-source community: StreamMapNet, MapTR, MapUncertaintyPrediction .

Citation

If the paper and code help your research, please kindly cite:

@inproceedings{liu2025uncertainty,
  title={Uncertainty-Instructed Structure Injection for Generalizable HD Map Construction},
  author={Liu, Xiaolu and Yang, Ruizi and Wang, Song and Li, Wentong and Chen, Junbo and Zhu, Jianke},
  booktitle={Proceedings of the Computer Vision and Pattern Recognition Conference},
  pages={22359--22368},
  year={2025}
}

​