This repository is dedicated to the implementation of the SAR Distorted Image translator Network (SARDINet). This neural network aims at translating SAR distorted images into optical ones.

One can find a detailed description of the network and the results in Deep Learning of Radiometrical and Geometrical Sar Distorsions for Image Modality translations.

The code is the baseline to compute the results obtained in the paper.

Data used in the paper are available on IEEE Dataport :

Abdourrahmane ATTO, January 9, 2022, "Can Artificial Intelligence Untangle Distorted and Compressed Geometries Associated with SAR Images of 3D Objects ? ", IEEE Dataport, doi: https://dx.doi.org/10.21227/y3pm-1113.

The code was computed using Python 3.10.6 and the packages versions detailed in the file requirements.txt.

In order to run the code please follow the next steps :

• Open the main.py file
• Choose the path where your data are located and the path where you want to save the results
• Choose you hyperparameters, whether to compute an adversarial training and your loss functions
• If you want to modify the architecture of the network, you'll find all you need in the TransNet.py file.
• Save your changes
• Run the main.py file

If this work was useful for you, please ensure citing our works :

BRALET, Antoine, ATTO, Abdourrahmane M., CHANUSSOT, Jocelyn, et al. Deep Learning of Radiometrical and Geometrical Sar Distorsions for Image Modality translations. In : 2022 IEEE International Conference on Image Processing (ICIP). IEEE, 2022. p. 1766-1770.

Thank you for your support

If you have any troubles with the article or the code, do not hesitate to contact us !