# CMOS

**Repository Path**: csc105/CMOS

## Basic Information

- **Project Name**: CMOS
- **Description**: This repository is the official PyTorch implementation of Better “CMOS” Produces Clearer Images: Learning Space-Variant Blur Estimation for Blind Image Super-Resolution
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2024-01-11
- **Last Updated**: 2024-06-12

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README


# Better “CMOS” Produces Clearer Images: Learning Space-Variant Blur Estimation for Blind Image Super-Resolution (CMOS, CVPR 2023)

This repository is the official PyTorch implementation of Better “CMOS” Produces Clearer Images: Learning Space-Variant Blur Estimation for Blind Image Super-Resolution
([arxiv](https://arxiv.org/abs/2304.03542)). 
Please feel free to contact us if you have any questions.


## Requirements
- Prepare experimental environments
  ```
  pip install -r requirements.txt
  ```

Note: this repository is based on [BasicSR](https://github.com/xinntao/BasicSR#memo-codebase-designs-and-conventions) and [MANet](https://github.com/JingyunLiang/MANet). Please refer to their repository for a better understanding of the code framework.


## Data Preparation
To prepare data, download datasets [NYUv2-BSR](https://drive.google.com/file/d/1W9zy45nvje8zQ7QaU0DK1c_9hPeGqB8R/view?usp=drive_link) and [Cityscapes-BSR](https://drive.google.com/file/d/1RLcLZbdq7qhqDgl4elywKSZX3Sq7Nwd3/view?usp=drive_link) from Google Drive and place them in the './datasets' folder. The creation of these two datasets is based on [this repository](https://github.com/codeslake/SYNDOF). Please refer to our paper for more detais.


## Training
Step1: to train CMOS with NYUv2-BSR, run this command:

```
python train.py --opt options/train/NYUv2_BSR/train_stage1.yml
```

Step2: to train non-blind RRDB-SFT with NYUv2-BSR, run this command:

```
python train.py --opt options/train/NYUv2_BSR/train_stage2.yml
```

Step3: to fine-tune RRDB-SFT with CMOS, run this command:

```bash
python train.py --opt options/train/NYUv2_BSR/train_stage3.yml
```

All trained models are provided in `./codes/pretrained_models`.


## Testing

To test CMOS (stage1, blur and semantic estimation) with NYUv2-BSR, run this command:

```
python test.py --opt options/test/NYUv2_BSR/test_stage1.yml
```
To test RRDB-SFT (stage2, non-blind SR with ground-truth blur and semantic maps) with NYUv2-BSR, run this command:

```
python test.py --opt options/test/NYUv2_BSR/test_stage2.yml
```
To test CMOS+RRDB-SFT (stage3, blind SR), run this command:

```
python test.py --opt options/test/NYUv2_BSR/test_stage3.yml
```

## Citation
    @inproceedings{chen2023better,
        title={Better" CMOS" Produces Clearer Images: Learning Space-Variant Blur Estimation for Blind Image Super-Resolution},
        author={Chen, Xuhai and Zhang, Jiangning and Xu, Chao and Wang, Yabiao and Wang, Chengjie and Liu, Yong},
        booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
        pages={1651--1661},
        year={2023}
    }

## Acknowledgement

The codes are based on [BasicSR](https://github.com/xinntao/BasicSR), [MANet](https://github.com/JingyunLiang/MANet) and [MTI-Net](https://github.com/SimonVandenhende/Multi-Task-Learning-PyTorch). Thanks for their great works.

## Reproduction

Name: Liming Xu
Time: 2023-11-18
Review: The network structure used in this paper is relatively complex, resulting in lower efficiency. There are misalignments between semantics and blur information, and solutions targeting this issue need to be proposed for improvement.