InnovateMU Hackathon

Malaria Image Classification using CNN

A Convolutional Neural Network (CNN) based machine learning model to detect malaria-infected blood cells. This project was developed for the InnovateMU Hackathon to demonstrate how Machine Learning can assist in early malaria diagnosis.

Project Overview

Malaria is a life-threatening disease affecting millions worldwide. Early and accurate detection is critical but often limited by resource constraints and reliance on manual microscopy. This project automates malaria detection using cell image classification, providing fast, reliable predictions.

Key features:

• Classifies blood cell images as Parasitized or Uninfected.
• CNN-based model trained on the Kaggle Malaria Dataset.
• High validation accuracy (~94%).
• Interactive Streamlit App for image upload and instant predictions.

Getting Started

Prerequisites

Make sure you have Python 3.x installed. Install required packages:

pip install -r requirements.txt

Run the Streamlit App

Launch the web app to test predictions:

streamlit run app.py

Dataset

The model uses the Cell Images for Detecting Malaria dataset from Kaggle.

• Contains images of Parasitized and Uninfected blood cells.
• Images are resized to 128x128 pixels for model training.

Model

• Architecture: CNN with Conv2D, MaxPooling, Dense layers, and Dropout
• Loss Function: Binary cross-entropy
• Optimizer: Adam
• Performance: Validation accuracy ~94%, with evaluation metrics including confusion matrix and classification report.

Usage

1. Clone the repository:

git clone <repository-url>
cd InnovateMUMachineLearningModel

2. Install dependencies:

pip install -r requirements.txt

3. Run the Streamlit app:

streamlit run app.py

4. Upload an image of a blood cell and see the prediction (Parasitized vs Uninfected).

Example Outputs

Acknowledgements

• Dataset: Kaggle - Cell Images for Detecting Malaria
• Hackathon: InnovateMU 2025