ESRGAN in tensorflow

Enhanced Super Resolution Generative Adversarial Network in tensorflow

This repo is based on pytorch impl original here

Work In Process :)

Requirements

• python 2.x / 3.x
• tensorflow-gpu 1.x
• opencv
• glob
• tqdm

Repo-Tree

│
├── output  (generated images)
│     ├── ...
│     └── xxx.png
├── tb_logs (tensorboard records)
│     ├── [unique id]
│     │     ├── *.ckpt
│     │     ├── *.tsv
│     │     ├── *.meta
│     │     └── ...
│     └── [unique id]
├── requirements.txt  (requirements)
├── readme.md         (explaination)
├── losses.py         (useful losses)
├── metrics.py        (useful metrics)
├── model.py          (ESRGAN model)
├── main.py           (trainer / inferener)
├── config.py         (global configurations)
├── tfutils.py        (useful TF utils)
├── utils.py          (image processing utils)
└── dataloader.py     (DataSet loader)

Usage

1. Clone this github repo.

git clone https://github.com/kozistr/ESRGAN-tensorflow
cd ESRGAN-tensorflow

2. install required packages (if needed)

# with pip
python -m pip install -r requirements.txt

# with conda
conda install --yes --file requirements.txt

3. run scripts!

For training,

python3 train.py

For evaluation,

python3 evaluate.py

For inference,

python3 inference.py --src test-lr.png --dst test-hr.png

Results

Citation

@InProceedings{wang2018esrgan,
    author = {Wang, Xintao and Yu, Ke and Wu, Shixiang and Gu, Jinjin and Liu, Yihao and Dong, Chao and Qiao, Yu and Loy, Chen Change},
    title = {ESRGAN: Enhanced super-resolution generative adversarial networks},
    booktitle = {The European Conference on Computer Vision Workshops (ECCVW)},
    month = {September},
    year = {2018}
}

Author

HyeongChan kim / kozistr