Aryog/Medical_Image_SuperResolution_SRGAN

Minor Project for IOE Thapathali Campus

MedSRGAN

PyTorch implementation of "MedSRGAN: medical images super-resolution using generative adversarial networks"

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:

   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:

   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.