Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/chaofengc/femasr

PyTorch codes for "Real-World Blind Super-Resolution via Feature Matching with Implicit High-Resolution Priors", ACM MM2022 (Oral)
https://github.com/chaofengc/femasr

blind-image-super-resolution real-image-super-resolution super-resolution vqgan

Last synced: about 8 hours ago
JSON representation

PyTorch codes for "Real-World Blind Super-Resolution via Feature Matching with Implicit High-Resolution Priors", ACM MM2022 (Oral)

Host: GitHub
URL: https://github.com/chaofengc/femasr
Owner: chaofengc
License: other
Created: 2022-01-05T07:13:32.000Z (almost 3 years ago)
Default Branch: main
Last Pushed: 2023-10-18T13:23:54.000Z (about 1 year ago)
Last Synced: 2024-12-26T19:07:34.270Z (7 days ago)
Topics: blind-image-super-resolution, real-image-super-resolution, super-resolution, vqgan
Language: Python
Homepage:
Size: 167 MB
Stars: 213
Watchers: 5
Forks: 15
Open Issues: 7
Metadata Files:
- Readme: README.md
- License: LICENSE

Awesome Lists containing this project

README

        # FeMaSR 

This is the official PyTorch codes for the paper  

[Real-World Blind Super-Resolution via Feature Matching with Implicit High-Resolution Priors (MM22 Oral)](https://arxiv.org/abs/2202.13142)  

[Chaofeng Chen\*](https://chaofengc.github.io), [Xinyu Shi\*](https://github.com/Xinyu-Shi), [Yipeng Qin](http://yipengqin.github.io/), [Xiaoming Li](https://csxmli2016.github.io/), [Xiaoguang Han](https://mypage.cuhk.edu.cn/academics/hanxiaoguang/), [Tao Yang](https://github.com/yangxy), [Shihui Guo](http://guoshihui.net/)   

(\* indicates equal contribution)

[![arXiv](https://img.shields.io/badge/arXiv-Paper-.svg)](https://arxiv.org/abs/2202.13142)

 

[![wandb](https://img.shields.io/badge/Weights_&_Biases-FFBE00?&logo=WeightsAndBiases&logoColor=white)](https://wandb.ai/chaofeng/FeMaSR?workspace=user-chaofeng)

![visitors](https://visitor-badge.laobi.icu/badge?page_id=chaofengc/FeMaSR)

[![LICENSE](https://img.shields.io/badge/LICENSE-CC%20BY--NC--SA%204.0-lightgrey)](https://github.com/chaofengc/FeMaSR/blob/main/LICENSE)

![framework_img](framework_overview.png)

### Update

- **2022.10.10** Release reproduce training log for SR stage in [![wandb](https://img.shields.io/badge/Weights_&_Biases-FFBE00?&logo=WeightsAndBiases&logoColor=white)](https://wandb.ai/chaofeng/FeMaSR?workspace=user-chaofeng). Reach similar performance as the paper, `LPIPS: 0.329 @415k` for div2k (x4).

- **2022.09.26** Add example training log with 70k iterations [![wandb](https://img.shields.io/badge/Weights_&_Biases-FFBE00?&logo=WeightsAndBiases&logoColor=white)](https://wandb.ai/chaofeng/FeMaSR?workspace=user-chaofeng)

- **2022.09.23** Add colab demo   

- **2022.07.02**  

      - Update codes of the new version FeMaSR   

      - Please find the old QuanTexSR in the `quantexsr` branch

Here are some example results on test images from [BSRGAN](https://github.com/cszn/BSRGAN) and [RealESRGAN](https://github.com/xinntao/Real-ESRGAN).

---

**Left**: [real images](./testset) **|** **Right**: [super-resolved images with scale factor 4](./results)

 

 

 

 

 

## Dependencies and Installation

- Ubuntu >= 18.04

- CUDA >= 11.0

- Other required packages in `requirements.txt`

```

# git clone this repository

git clone https://github.com/chaofengc/FeMaSR.git

cd FeMaSR 

# create new anaconda env

conda create -n femasr python=3.8

source activate femasr 

# install python dependencies

pip3 install -r requirements.txt

python setup.py develop

```

## Quick Inference

```

python inference_femasr.py -s 4 -i ./testset -o results_x4/

python inference_femasr.py -s 2 -i ./testset -o results_x2/

```

## Train the model

### Preparation

#### Dataset

Please prepare the training and testing data follow descriptions in the main paper and supplementary material. In brief, you need to crop 512 x 512 high resolution patches, and generate the low resolution patches with [`degradation_bsrgan`](https://github.com/cszn/BSRGAN/blob/3a958f40a9a24e8b81c3cb1960f05b0e91f1b421/utils/utils_blindsr.py?_pjax=%23js-repo-pjax-container%2C%20div%5Bitemtype%3D%22http%3A%2F%2Fschema.org%2FSoftwareSourceCode%22%5D%20main%2C%20%5Bdata-pjax-container%5D#L432) function provided by [BSRGAN](https://github.com/cszn/BSRGAN). While the synthetic testing LR images are generated by the [`degradation_bsrgan_plus`](https://github.com/cszn/BSRGAN/blob/3a958f40a9a24e8b81c3cb1960f05b0e91f1b421/utils/utils_blindsr.py?_pjax=%23js-repo-pjax-container%2C%20div%5Bitemtype%3D%22http%3A%2F%2Fschema.org%2FSoftwareSourceCode%22%5D%20main%2C%20%5Bdata-pjax-container%5D#L524) function for fair comparison.

#### Model preparation

Before training, you need to

- Download the pretrained HRP model: [generator](https://github.com/chaofengc/FeMaSR/releases/download/v0.1-pretrain_models/FeMaSR_HRP_model_g.pth), [discriminator](https://github.com/chaofengc/FeMaSR/releases/download/v0.1-pretrain_models/FeMaSR_HRP_model_d.pth) 

- Put the pretrained models in `experiments/pretrained_models`

- Specify their path in the corresponding option file.

### Train SR model

```

python basicsr/train.py -opt options/train_FeMaSR_LQ_stage.yml

```

### Model pretrain

In case you want to pretrain your own HRP model, we also provide the training option file:

```

python basicsr/train.py -opt options/train_FeMaSR_HQ_pretrain_stage.yml

```

## Citation

```

@inproceedings{chen2022femasr,

      author={Chaofeng Chen and Xinyu Shi and Yipeng Qin and Xiaoming Li and Xiaoguang Han and Tao Yang and Shihui Guo},

      title={Real-World Blind Super-Resolution via Feature Matching with Implicit High-Resolution Priors}, 

      year={2022},

      Journal = {ACM International Conference on Multimedia},

}

```

## License


This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

## Acknowledgement

This project is based on [BasicSR](https://github.com/xinntao/BasicSR).