Under-scanning_NLOS

Deep Non-line-of-sight Imaging from Under-scanning Measurements (NeurIPS 2023)

Pipeline

Reconstructed Results from Real-world Measurements Captured by FK.

Reconstructed Results from Real-world Measurements Captured by NLOST.

Trade-off about the Sampling Points and Noise Levels

Datasets

Synthetic Data
We utilized the synthetic data (~3000 motorbike dataset) provided by LFE.
You can download Here

Real-world Data
We utilized the real-world data provided by FK and NLOST.
Or you can download the preprocessed data Here.

Experiment

We first train the network on Synthetic Data, and then directly test on Real-world Data.

Training

Modify the data path and then run bash train.sh

Testing

1, [Synthetic] Modify the data path and then run python validate_syn.py
2, [Real-world] Modify the data path and then run python validate_fk.py and python validate_cvpr2023.py

Contact

For questions, feel free to contact us (yueli65@mail.ustc.edu.cn).

Acknowledgements

We thank the authors who shared the code of their works. Particularly LFE.

Citation

If you find it useful, please cite our paper.

@inproceedings{li2023deep,
title={Deep Non-line-of-sight Imaging from Under-scanning Measurements},
author={Li, Yue and Zhang, Yueyi and Ye, Juntian and Xu, Feihu and Xiong, Zhiwei},
booktitle={Thirty-seventh Conference on Neural Information Processing Systems},
year={2023}
}