# xinntao/Real-ESRGAN

**Attribution required: if you use, quote, or summarise this content, you must credit and link back to [awesome-repositories.com](https://awesome-repositories.com/repository/xinntao-real-esrgan).**

34,385 stars · 4,266 forks · Python · bsd-3-clause

## Links

- GitHub: https://github.com/xinntao/Real-ESRGAN
- awesome-repositories: https://awesome-repositories.com/repository/xinntao-real-esrgan.md

## Topics

`amine` `denoise` `esrgan` `image-restoration` `jpeg-compression` `pytorch` `real-esrgan` `super-resolution`

## Description

Real-ESRGAN is a deep learning restoration pipeline designed to enhance low-resolution media and improve the visual quality of damaged photographs. It functions as a generative image upscaler that reconstructs high-resolution details from source inputs by utilizing neural networks trained to fill in missing information and remove noise.

The project distinguishes itself as a blind super-resolution tool, meaning it improves image sharpness and fidelity without requiring prior knowledge of the specific degradation applied to the source. It employs high-order degradation modeling to address complex, real-world artifacts and utilizes a generative adversarial network architecture to refine output realism. By applying these techniques, the system effectively increases pixel density while preserving sharp edges and textures.

The software supports a range of media upscaling workflows, including memory-efficient tiled processing for handling large images. It provides a framework for computer vision preprocessing and legacy content archiving, allowing users to execute pre-trained weight inference to transform input pixels into clearer, high-definition outputs.

## Tags

### Graphics & Multimedia

- [Image Super-Resolution](https://awesome-repositories.com/f/graphics-multimedia/image-super-resolution.md) — Enhances low-resolution media files by applying machine learning models to sharpen visual details and increase pixel density. ([source](https://github.com/xinntao/Real-ESRGAN))
- [Image Restoration](https://awesome-repositories.com/f/graphics-multimedia/image-restoration.md) — Improves the visual quality of damaged or low-resolution photographs by automatically filling in missing details and removing noise.
- [Media Upscaling Workflows](https://awesome-repositories.com/f/graphics-multimedia/media-upscaling-workflows.md) — Increases the resolution of digital images and videos while preserving sharp edges and textures.

### Artificial Intelligence & ML

- [Generative Upscalers](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-upscalers.md) — Reconstructs high-resolution details from low-resolution input images using deep neural networks trained on synthetic degradation processes.
- [Blind Restoration Models](https://awesome-repositories.com/f/artificial-intelligence-ml/blind-restoration-models.md) — Improves visual quality and sharpness in images without requiring prior knowledge of the specific degradation applied.
- [Generative Adversarial Networks](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-adversarial-networks.md) — Implements a dual-model architecture where a generator and discriminator refine image quality through adversarial training.
- [Convolutional Neural Networks](https://awesome-repositories.com/f/artificial-intelligence-ml/convolutional-neural-networks.md) — Uses stacked convolutional layers to extract spatial features and reconstruct fine details from low-resolution inputs.
- [Inference Engines](https://awesome-repositories.com/f/artificial-intelligence-ml/inference-engines.md) — Executes optimized mathematical operations using static model parameters to transform input pixels into enhanced output.
- [Degradation Modeling](https://awesome-repositories.com/f/artificial-intelligence-ml/degradation-modeling.md) — Applies multiple rounds of image corruption to simulate and handle complex, real-world artifacts.