Medical_Image_SuperResolution_SRGAN
所属分类:生物医药技术
开发工具:Python
文件大小:0KB
下载次数:0
上传日期:2024-01-24 17:57:11
上 传 者:
sh-1993
说明: IOE Thapathali校区小项目
(Minor Project for IOE Thapathali Campus)
文件列表:
__pycache__/
custom_dataset/
img/
check.jpeg
dataset.py
discriminator.py
enhanced_output.jpeg
generator.py
losses.py
main.py
mode.py
tester.py
vgg19.py
# MedSRGAN
## PyTorch implementation of "MedSRGAN: medical images super-resolution using generative adversarial networks"
```python
import torch
from generator import Generator
from discriminator import Discriminator
generator = Generator(
in_channels= 3,
blocks= 8
)
discriminator = Discriminator(
in_channels= 3,
img_size= (256, 256)
)
```
### **Using the App**
To use the app, follow these steps:
1. Create the **`custom_dataset`** folder in your project directory.
2. Create the **`train_LR`** and **`train_HR`** subdirectories inside **`custom_dataset`**
3. Run the following command in the terminal to train the model:
```bash
python main.py --LR_path custom_dataset/train_LR --GT_path custom_dataset/train_HR
```
This will train the MedSRGAN model using your medical image dataset. Adjust the hyperparameters in the **`main.py`** file as needed.
4. After training, you can test the model on new images using:
```bash
python tester.py
```
Make sure to input the path of the test image when prompted.
5. View the output result as **`enhanced_output.jpeg`** in your **root** directory.
近期下载者:
相关文件:
收藏者: