### Train FBCNN Model

Source: https://github.com/jiaxi-jiang/fbcnn/blob/main/README.md

Run this command to start the training process for the FBCNN model. The configuration file is located in the './options/' directory.

```bash
python main_train_fbcnn.py
```

--------------------------------

### Test FBCNN on Real-World JPEG Images

Source: https://github.com/jiaxi-jiang/fbcnn/blob/main/README.md

Use this command to test the FBCNN model on real-world JPEG images.

```bash
python main_test_fbcnn_color_real.py
```

--------------------------------

### Test FBCNN on Grayscale JPEG Images

Source: https://github.com/jiaxi-jiang/fbcnn/blob/main/README.md

Use this command to test the FBCNN model on grayscale JPEG images.

```bash
python main_test_fbcnn_gray.py
```

--------------------------------

### Test FBCNN on Color JPEG Images

Source: https://github.com/jiaxi-jiang/fbcnn/blob/main/README.md

Use this command to test the FBCNN model on color JPEG images.

```bash
python main_test_fbcnn_color.py
```

--------------------------------

### Test FBCNN on Grayscale JPEG Images (Double JPEG Degradation)

Source: https://github.com/jiaxi-jiang/fbcnn/blob/main/README.md

Use this command to test the FBCNN model on grayscale JPEG images that have been compressed twice.

```bash
python main_test_fbcnn_gray_doublejpeg.py
```

--------------------------------

### Double JPEG Degradation Model

Source: https://github.com/jiaxi-jiang/fbcnn/blob/main/README.md

This model describes the process of non-aligned double JPEG compression, where an image is compressed twice with potential shifts. This degradation model can be integrated into other image restoration tasks.

```mathematics
y = JPEG(shift(JPEG(x, QF1)),QF2)
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.