# 无参考图像质量评价

**Repository Path**: buptybx/NRVQA

## Basic Information

- **Project Name**: 无参考图像质量评价
- **Description**: 无参考图片/视频质量评价(No-Reference Blind Video Quality Assessment)
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 5
- **Created**: 2023-02-08
- **Last Updated**: 2024-06-23

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# 无参考图片/视频质量评价(No-Reference Blind Video Quality Assessment)

## 深度学习

- [deepbiq](https://github.com/zhl2007/pytorch-image-quality-param-ctrl)
- [DIQA](https://towardsdatascience.com/deep-image-quality-assessment-with-tensorflow-2-0-69ed8c32f195)
- [VSFA](https://github.com/lidq92/VSFA)

## 传统方法

- BRISQUE(**extract 36 dimesion brisque features,you can train svr model in labeled datasets like TID2013/LIVE/CSIQ**)
- NIQE
- PIQE


## test

### brisque
*high score has high quality*
```
python test.py --mode brisque --path=imgs/origin.jpeg
python test.py --mode brisque --path=imgs/compression.jpeg
```
### niqe
*high score has low quality*
```
python test.py --mode niqe --path=imgs/origin.jpeg
python test.py --mode niqe --path=imgs/compression.jpeg
```
### piqe
*high score has low quality*
```
python test.py --mode piqe --path=imgs/origin.jpeg
python test.py --mode piqe --path=imgs/compression.jpeg
```
  




## 相关论文[reference papers](https://github.com/buyizhiyou/papers/tree/master/VQA_IQA)

## Thanks for your star!