African Deep Tech Hackathon – Malaria Detection App
Features
- Upload cell images and detect malaria infection (Parasitized or Uninfected)
- Medical recommendation and risk level output
- API access for integration
- Docker support for easy deployment
Getting Started
1. Clone the Repository
git clone
github.com
cd african-deep-tech-hackathon
3. Install Requirements (Python 3.8+)
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
pip install -r requirements.txt
Place your trained YOLOv8 classification model at:
runs/classify/train4/weights/best.pt
You can change the path in the MODEL_PATH variable in app.py if needed.
4. Run the App
python app.py
Visit
localhost to use the web interface.
Using Docker
1. Build Docker Image
docker build -t malaria-hackathon:latest .
2. Run the Container
docker run -d -p 5000:5000 --name malaria-app malaria-hackathon:latest
Visit
localhost
API Endpoint
POST /api/predict
Form Data:
- file: image file (.jpg, .png, etc.)
Response Example:
{
"prediction": "Parasitized",
"confidence": 0.98,
"all_probabilities": {
"Parasitized": 0.98,
"Uninfected": 0.02
},
"recommendation": "...",
"risk_level": "danger"
}
Sample Test Image
You can use test cell images from the Cell Images for Detecting Malaria dataset: