elishaaminu90/malaria-scan-cnn

🔬 MalariaScan — Malaria Cell Classifier

A Convolutional Neural Network (CNN) that detects malaria-infected vs uninfected blood cells from microscope images.

Model Performance

Achieves 94% overall accuracy on the validation set:

Class	Precision	Recall	F1-Score
Parasitized (Infected)	0.97	0.91	0.94
Uninfected	0.92	0.97	0.94

Model Architecture

• 3 Conv2D + MaxPooling blocks (16 → 32 → 64 filters)
• Dropout regularization (0.2 - 0.5)
• Dense layer (64 units) + sigmoid output
• 826,529 total parameters

Live Demo

🔗 Try it live on Streamlit

Dataset

Trained on the Cell Images for Detecting Malaria dataset (NIH).

Usage

from tensorflow.keras.models import load_model
model = load_model('malaria_cnn_model.h5')

GitHub

Full code: https://github.com/elishaaminu90-debug/Cell-Vision.AI

Author

Elisha Aminu — GitHub | LinkedIn

Disclaimer

For educational purposes only. Not a substitute for professional medical diagnosis.