U-Net based neural network for fringe pattern denoising

This repository is for the LRDUNet model proposed in the following paper:

Javier Gurrola-Ramos, Oscar Dalmau and Teresa E. Alarcón, "U-Net based neural network for fringe pattern denoising", Optics and Lasers in Engineering, vol 149, pp. 106829, 2022, doi: 10.1016/j.optlaseng.2021.106829.

If you use this paper work in your research or work, please cite our paper:

@article{gurrola2022u,
  title = {U-Net based neural network for fringe pattern denoising},
  journal = {Optics and Lasers in Engineering},
  volume = {149},
  pages = {106829},
  year = {2022},
  issn = {0143-8166},
  doi = {https://doi.org/10.1016/j.optlaseng.2021.106829},
  url = {https://www.sciencedirect.com/science/article/pii/S0143816621002992},
  author = {Javier Gurrola-Ramos and Oscar Dalmau and Teresa Alarcón},
}

Link to download the pretrained model, and the training and test datasets.

• Python 3.6
• Numpy 1.19.2
• PyTorch 1.8.0
• torchvision 0.9.0
• pytorch-msssim 0.2.1
• ptflops 0.6.4
• tqdm 4.49.0
• scikit-image 0.17.2
• sewar 0.4.4
• yaml 0.2.5

Default parameters used in the paper are set in the config.yaml file:

groups: 8
dense covolutions: 2
downsampling: 'Conv2'
residual: true
patch size: 64
batch size: 16
learning rate: 1.e-3
weight decay: 1.e-2
scheduler gamma: 0.5
scheduler step size: 5
epochs: 20

Additionally, you can choose the device, the number of workers of the data loader, and enable multiple GPU use.

To train the model use the following command:

python main_train.py

Place the pretrained model in the './Pretrained' folder. To test the model in the simulated fringe patterns, use the following command:

python main_test_simulated.py

To test the model in the experimetal fringe patterns, use the following command:

python main_test_experimental.py

If you have any question about the code or paper, please contact [email protected] .