NOTE: In case you are not able to view the ipynb files on github, then paste its link on nbviewer, https://nbviewer.jupyter.org/

In this repo I have implemented three different deep learning architectures for image denoising,
REDNet; https://arxiv.org/pdf/1606.08921.pdf
Multi-level Wavelet CNN (MWCNN); https://arxiv.org/pdf/1805.07071.pdf
PRIDNet; https://arxiv.org/pdf/1908.00273.pdf

I found that PRIDNet was giving the best results as compared to the other two, giving me the best PSNR and SSIM scores along with best visualizations which were very close to the ground truth images.

Notice the level of detail in the eye ball.