This preject implemented SR-GAN algorithm on super resolution. Paper link: https://arxiv.org/pdf/1609.04802.pdf
The result like this:
It's very impressive and do every vivid detail super resolution on low quality images. For now, the pretrained weights is not supported.
To run this repo, you ganna need install some dependencies.
tensorflow-2.0
alfred-py
numpy
matplotlib
opencv-python
We support training on DIV2K dataset, raw data can be obtained from: https://data.vision.ee.ethz.ch/cvl/DIV2K/
After download train/validate images (both HR and LR), soft link them to ./div2k
:
Then, using:
python3 train_srgan.py