Super Resolution of Medical X-Ray Scans

I have implemented the Super Resolution Convolutional Neural Network model on custom X-Ray Scans dataset. The model achieved 42.47 PSNR on the test set. I have also made an app using StreamLit.

Dataset

You can download the train and test data for personal use. I have not provided a training file because the main purpose is to have an app that can aid doctors by providing X2 zoomed high resolution X-Ray scans.

Type	Link
Train set	Download
Test set	Download

Requirements

numpy
torch==1.7.1+cpu
streamlit==0.77.0
Pillow==8.1.1

# Install using pip
pip install -r requirements.txt

Running the App

Open the Root directory and run the following command in your terminal.

streamlit run app.py

This will output the URL, open it in your browser. You can then upload the X-ray Scan and get the SRCNN X2 output image.