Lei Yang · Xinyu Zhang · Jun Li · Li Wang · Chuang Zhang · Li Ju · Zhiwei Li · Yang Shen
SGV3D is an innovative scenario generalization framework for vision-based roadside 3D object detection. SGV3D, with only a minimal increase in latency, significantly outperforms other leading detectors by substantial margins of +42.57%, +5.87%, and +14.89% for three categories in DARI-V2X-I heterologous settings.
- [2025/03/07] Both arXiv and codebase are released!
This project is not possible without the following codebases.
If you use BEVHeight in your research, please cite our work by using the following BibTeX entry:
@article{yang2024sgv3d,
title={SGV3D: Towards scenario generalization for vision-based roadside 3D object detection},
author={Yang, Lei and Zhang, Xinyu and Li, Jun and Wang, Li and Zhang, Chuang and Ju, Li and Li, Zhiwei and Shen, Yang},
journal={arXiv preprint arXiv:2401.16110},
year={2024}
}