https://github.com/zjbthomas/tmo-gan
https://github.com/zjbthomas/tmo-gan
Last synced: 9 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/zjbthomas/tmo-gan
- Owner: zjbthomas
- Created: 2022-05-26T02:15:01.000Z (about 4 years ago)
- Default Branch: main
- Last Pushed: 2023-04-03T15:30:23.000Z (about 3 years ago)
- Last Synced: 2025-10-09T10:07:23.779Z (9 months ago)
- Language: Python
- Size: 73.2 MB
- Stars: 4
- Watchers: 1
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# A Generative Adversarial Network Based Tone Mapping Operator for 4K HDR Images
This is the repository for paper [A Generative Adversarial Network Based Tone Mapping Operator for 4K HDR Images](https://ieeexplore.ieee.org/abstract/document/10074176/) accepted to ICNC 2023.
## Dataset
The dataset we constructed in Section 2.1 of our paper can be downloaded [here](https://www.dropbox.com/s/0nh4837okzd0jtq/data_train.tar.xz?dl=0).
## Experiments
### Training
Run the following code to train the network:
```
python -u esrgan.py --hdr_image_dir /path/to/HDR/ --sdr_image_dir /path/to/SDR/ --residual_blocks 23 --lr 1e-5 --warmup_epochs 300 --image_size 128 --batch_size 8 --psnr_decay_epoch 50 --decay_epoch 100 --n_epochs 600
```
### Testing
Run the following code to evaluate the network. A pretrained model is provided [here](./pretrained/generator.pth).
```
python -u infer.py --image_path /path/to/HDR/dir/
--output_dir /path/to/output/dir/ --checkpoint_model ./pretrained/generator.pth
```