Complex-valued Wasserstein GAN for Sentinel-1 Image Superresolution

Complex-WGAN network adapted from https://github.com/jeremyfix/complex-wgan.git and modified for Sentinel-1 superresolution purpose. This repository provides the inference code as well as pretrained networks for the complex-wgan paper.

To test inferences, you can proceed by

1- creating a virtual environment

cd Code
python3 -m venv venv
source venv/bin/activate

2- Install the required dependencies

python -m pip install -r requirements.txt

3- Download Sentinel-1 data from Copernicus (adjust start_date, end_date as needed)

python S1_download.py

4- Unzip, patch up and downsample to create input data (toggle action, patch_making for different needs)

python S1_process.py

5- Train the model with new SAR data

python train.py --epochs 200

6- Execute the sampling and interpolation script

python test_sample_S1.py --load_model Sentinel-1 

7- Computes the FIDs

python utils/fid.py  --load_model Sentinel-1 --postprocess "None" --fold train
python utils/fid.py  --load_model Sentinel-1 --postprocess "None" --fold test

Example interpolations

On SAR

Left Interpolation in the latent space given three independent latent vectors $z_0, z_1, z_2$. The generated samples associated with $z_0$, $z_1$ and $z_2$ are respectively in the top left, top right and bottom left corners.

Right Interpolation in the latent space by applying a rotation of a random latent vector $z_0$.

For all these representations, the complex numbers are represented in the HSV colorspace.