This repository contains the implementation of an eye disease classification system using Convolutional Neural Networks (CNN), VGG16, and ResNet-50 architectures. The project aims to accurately classify various eye diseases from retinal images.
This project is part of the GirlScript Summer of Code 2024 program.
- Overview
- Table of Contents
- Features
- Dataset
- Requirements
- Installation
- Usage
- Model Architecture
- Training
- License
- Data preprocessing and augmentation.
- Implementation of CNN, VGG16, and ResNet-50 models.
- Training and validation scripts.
The dataset used for this project consists of retinal images labeled with various eye diseases. You can download the dataset from Kaggle Eye Disease Dataset.
- Python 3.7+
- TensorFlow 2.x
- Keras
- NumPy
- Pandas
- Matplotlib
- scikit-learn
-
Clone the repository:
git clone https://github.com/yourusername/eye-disease-classification.git cd eye-disease-classification
-
Install the required packages:
pip install -r requirements.txt
-
Preprocess the Data:
- Download the dataset and place it in the
data/
directory. - Run the preprocessing script to prepare the data for training:
python preprocess.py
- Download the dataset and place it in the
-
Train the Model:
- Train the CNN model:
python train.py --model cnn
- Train the VGG16 model:
python train.py --model vgg16
- Train the ResNet-50 model:
python train.py --model resnet50
- Train the CNN model:
-
Evaluate the Model:
- Evaluate the trained model:
python evaluate.py --model model_name
- Evaluate the trained model:
A custom CNN architecture designed for image classification.
A pre-trained VGG16 model fine-tuned for eye disease classification.
A pre-trained ResNet-50 model fine-tuned for eye disease classification.
The training process involves:
- Splitting the dataset into training, validation, and test sets.
- Data augmentation to improve model generalization.
- Training the model with early stopping and model checkpointing.
This project is licensed under the MIT License. See the LICENSE file for details.