Classification and Segmentation:

This project was tested on Ubuntu 22.04 LTS with Python 3.10

Summary:

There are 2 aspects to this project:

1. Classification of Diabetic Retinopathy in fundus images. Classification is done into 5 classes:
   • 0: No DR
   • 1: Mild
   • 2: Moderate
   • 3: Severe
   • 4: Proliferative DR
2. Segmentation of blood vessels in fundus images

How to run the program:

1. Install python 3.10

2. Install everything you need

   • python -m venv venv
   • Activate the virtual environment
     • Linux/MacOS: source venv/bin/activate
     • Windows: venv\Scripts\activate
   • pip install -r requirements.txt
   • To setup the GPU version of pytorch, follow the instructions in this link. A quick summary of the steps is given below:
     • pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 --no-cache-dir --force-reinstall
       • There should be a download size of 2.3GB if it is downloading the GPU version correctly. If it's something like 600MB, that's the CPU version.
       • If that didn't work, simply install the CPU version for now. It'll slow down transcription but it'll work.
         • pip install torch

3. python run.py

   • You can follow the on-screen instructions to run the program.
   • Input folder contains sample images to test the program. You can either feed the entire folder path or a single image path as input to the program.
   • run.py gives you a quick way to run the program, and is meant to be an example of how to use the various functions.

PS: Consider leaving a star if you find this useful. Thanks!

Display Items:

Confusion Matrix On Test Set:

More Samples: