Eye Disease Classifier

Eye Disease Classifier is a machine learning project designed to automatically detect and classify common eye diseases from retinal images.
Now features an interactive web-based UI built with Streamlit!

Live Demo

You can try the live demo of the Eye Disease Classifier here: Eye Disease Classifier Demo

Features

• Automatic Eye Disease Detection: Classifies retinal images into Cataract, Diabetic Retinopathy, Glaucoma, or Normal.
• Streamlit Web UI: User-friendly web interface for uploading images and viewing results instantly.
• Model Training and Evaluation: Easily train and evaluate models on your own dataset.
• Performance Visualization: Confusion matrices and results tracking.

Dataset

Kaggle: Eye Disease Retinal Images

Classes

• Cataract
• Diabetic Retinopathy
• Glaucoma
• Normal

Model Performance

• Test Accuracy: 0.900
• Test Loss: 0.387
• Epochs (Total): 61

Setup Instructions

Prerequisites

• Python 3.8 or higher
• pip (Python package installer)
• (Optional) CUDA-enabled GPU for faster model training/inference

Installation

1. Clone the repository:

   git clone https://github.com/TaigaTi/Eye-Disease-Classifier.git
   cd Eye-Disease-Classifier

2. Create and activate a virtual environment (recommended):

   python -m venv .venv
   # On Windows
   .venv\Scripts\activate
   # On macOS/Linux
   source .venv/bin/activate

3. Install required dependencies:

   pip install -r requirements.txt

Usage

Training the Model

python train_model.py

Running the Streamlit UI

Start the web application for interactive disease classification:

streamlit run app.py

Visit http://localhost:8501 in your browser.

Project Structure

Eye-Disease-Classifier/
│
├── models/                              # Trained models
├── confusion_matrices/                  # Confusion matrices
├── train_model.py                       # Trains a new model
├── app.py                               # Streamlit UI application
├── eye_classification_results.xlsx      # Results from model training
├── requirements.txt                     # Python dependencies
└── README.md                            # Project documentation