Superresolution based on ESCPN
Installation
pip3 install -r requirements.txt
Usage
Train the model(after put the train_HR images in folder images/):
python3 train.py
Superresolution(after put the test_LR images in folder test_images/):
python3 rebuild.py
Evaluation(after put the test_HR images(jpg) in the folder result/):
python3 psnr.py
Reference
Efficient Sub-Pixel Convolutional Neural Network(ESCPN)
https://arxiv.org/pdf/1609.05158.pdf
TensorFlow implementation of ESCPN
https://github.com/kweisamx/TensorFlow-ESPCN
https://github.com/Nyloner/TensorFlow-ESPCN