Self-supervised Depth Estimation from Thermal Image

This repository provides a integrated codebase for the following papers:

• Self-supervised Depth and Ego-motion Estimation for Monocular Thermal Video using Multi-spectral Consistency Loss (RA-L 21 & ICRA 22)
• Maximizing Self-supervision from Thermal Image for Effective Self-supervised Learning of Depth and Ego-motion (RA-L 22 & IROS 22)
• Self-supervised Monocular Depth Estimation from Thermal Images via Adversarial Multi-spectral Adaptation (WACV 23)

Updates

• 2023.03.30: Open Github page.
• TBA: The code will be released within one~two month.

License

Shield:

Our code is licensed under a MIT License.

Citation

Please cite at least one of the following papers if you use our work in your research.

@article{shin2021self,
  title={Self-supervised depth and ego-motion estimation for monocular thermal video using multi-spectral consistency loss},
  author={Shin, Ukcheol and Lee, Kyunghyun and Lee, Seokju and Kweon, In So},
  journal={IEEE Robotics and Automation Letters},
  volume={7},
  number={2},
  pages={1103--1110},
  year={2021},
  publisher={IEEE}
}

@article{shin2022maximizing,
  title={Maximizing Self-Supervision From Thermal Image for Effective Self-Supervised Learning of Depth and Ego-Motion},
  author={Shin, Ukcheol and Lee, Kyunghyun and Lee, Byeong-Uk and Kweon, In So},
  journal={IEEE Robotics and Automation Letters},
  volume={7},
  number={3},
  pages={7771--7778},
  year={2022},
  publisher={IEEE}
}

@inproceedings{shin2023self,
  title={Self-Supervised Monocular Depth Estimation From Thermal Images via Adversarial Multi-Spectral Adaptation},
  author={Shin, Ukcheol and Park, Kwanyong and Lee, Byeong-Uk and Lee, Kyunghyun and Kweon, In So},
  booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
  pages={5798--5807},
  year={2023}
}

Related projects

• ThermalSfMLearner-MS (RA-L 2021 & ICRA 2022)
• ThermalMonoDepth (RA-L 2022 & IROS 2022)