Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/leverxgroup/esrgan
Enhanced SRGAN. Champion PIRM Challenge on Perceptual Super-Resolution
https://github.com/leverxgroup/esrgan
computer-vision deep-neural-networks generative-adversarial-network image-processing pytorch super-resolution
Last synced: 2 months ago
JSON representation
Enhanced SRGAN. Champion PIRM Challenge on Perceptual Super-Resolution
- Host: GitHub
- URL: https://github.com/leverxgroup/esrgan
- Owner: leverxgroup
- License: other
- Created: 2020-10-16T07:13:46.000Z (over 3 years ago)
- Default Branch: master
- Last Pushed: 2022-09-17T10:32:31.000Z (almost 2 years ago)
- Last Synced: 2024-01-14T10:08:19.495Z (5 months ago)
- Topics: computer-vision, deep-neural-networks, generative-adversarial-network, image-processing, pytorch, super-resolution
- Language: Python
- Homepage: https://esrgan.readthedocs.io
- Size: 10.8 MB
- Stars: 121
- Watchers: 5
- Forks: 15
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Lists
- awesome-stars - leverxgroup/esrgan - Enhanced SRGAN. Champion PIRM Challenge on Perceptual Super-Resolution (Python)
README
# ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks
Pipeine for Image Super-Resolution task that based on a frequently cited paper, [ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks](https://arxiv.org/abs/1809.00219) (Wang Xintao et al.), published in 2018.
In few words, image super-resolution (SR) techniques reconstruct a higher-resolution (HR) image or sequence
from the observed lower-resolution (LR) images, e.g. upscaling of 720p image into 1080p.
One of the common approaches to solving this task is to use deep convolutional neural networks
capable of recovering HR images from LR ones. And ESRGAN (Enhanced SRGAN) is one of them.
Key points of ESRGAN:
- SRResNet-based architecture with residual-in-residual blocks;
- Mixture of context, perceptual, and adversarial losses. Context and perceptual losses are used for proper image upscaling,
while adversarial loss pushes neural network to the natural image manifold using a discriminator network
that is trained to differentiate between the super-resolved images and original photo-realistic images.
### Technologies
* `Catalyst` as pipeline runner for deep learning tasks. This new and rapidly developing [library](https://github.com/catalyst-team/catalyst).
can significantly reduce the amount of boilerplate code. If you are familiar with the TensorFlow ecosystem, you can think of Catalyst
as Keras for PyTorch. This framework is integrated with logging systems such as the well-known [TensorBoard](https://www.tensorflow.org/tensorboard);
* `Pytorch` and `torchvision` as main frameworks for deep learning;
* `Albumentations` and `PIQ` for data processing.
## Quick Start
### Setup environment
```bash
pip install git+https://github.com/leverxgroup/esrgan.git
```
### Run an experiment
```bash
catalyst-dl run -C esrgan/config.yml --benchmark
```
where `esrgan/config.yml` is a path to the [config](config.yml) file.
## Results
Some examples of work of ESRGAN model trained on [DIV2K](https://data.vision.ee.ethz.ch/cvl/DIV2K) dataset:
| LR(low resolution) | ESRGAN(original) | ESRGAN(ours) | HR(high resolution) |
|:---:|:---:|:---:|:---:|
| 
| 
| 
| 
|
| 
| 
| 
| 
|
| 
| 
| 
| 
|
## Documentation
Full documentation for the project is available at https://esrgan.readthedocs.io/
## License
`esrgan` is released under a CC BY-NC-ND 4.0 license. See [LICENSE](LICENSE) for additional details about it.