The Malaria Diagnostic Tool is a web application designed to assist in the detection of malaria using blood smear images. Leveraging a MobileNetV3Small-based Convolutional Neural Network (CNN), this tool provides an intuitive interface for uploading images and receiving instant diagnostic results.
- Image Upload: Users can upload blood smear images in PNG format.
- Instant Diagnosis: The tool processes the uploaded image and provides the likelihood of the sample being "Parasitized" or "Uninfected".
- User-Friendly Interface: A simple and clean design makes it easy for users to interact with the application.
- Probability Display: The tool shows the probability of each category, enhancing the transparency of the diagnosis.
- Backend: Flask for serving the application and handling API requests.
- Machine Learning Model: MobileNetV3Small-based CNN built with TensorFlow/Keras.
- Frontend: HTML, CSS, JavaScript, Tailwind CSS.
- Deployment: Render for hosting the backend server.
- Flask: The backend is built using Flask, a lightweight WSGI web application framework in Python.
- Model Loading: The MobileNetV3Small-based CNN model is loaded at startup and used for predictions.
- API Endpoint: A
/predictendpoint is provided to handle image uploads and return diagnostic results.
- HTML/CSS/JavaScript: The frontend is built with standard web technologies.
- Tailwind CSS: Used for styling to ensure a responsive and modern design.
- Image Upload: Implements drag-and-drop functionality and standard file input.
The application is deployed on Netlify at malaria-diagnosis.netlify.app.
This project is licensed under the MIT License - see the LICENSE file for details.
For any inquiries, please contact me at edwinmbonyjr@gmail.com or edwin.ade@stu.cu.edu.ng.