README.md

Purpose

Blur or restore images and videos using opencv and a Tensorflow/Keras implemented Generative Advesarial Network(GAN) model.

Requirements

numpy==1.20.3
tqdm==4.62.2
Keras==2.4.3
tensorflow==2.5.0
Pillow==8.3.1
opencv==4.5.1
Note: I recommend installing opencv through anaconda through the conda-forge channel.

References

Significant amount of code on GAN model building and data preprocessing borrowed from Rapael Meudec,
Github repo: https://github.com/RaphaelMeudec/deblur-gan
Upscale Models: https://github.com/Saafke/EDSR_Tensorflow.git