Logo Lanfrica

samehaisaa/TunisiaAI-Malaria-Detection

Domaine:

healthcare

Type de record:

model
Créateur:
sam
Hôte:
My internship project (May 26 – August 2024) creating a hybrid detection system (YOLOv8+SVM) for malaria diagnosis in Tunisian healthcare settings. It features cellular analysis via an interactive Streamlit interface, developed at CRISTAL Lab and in collaboration with Charles Nicole Hospital. # TunisiaAI Malaria Detection **This repository contains the work I developed during my summer internship at the CRISTAL Lab** *(Research Center in Networks, Image, Systems, Architecture, and Multimedia - LR99ES25)*. **The project was a collaborative effort between CRISTAL Lab and the Parasitology & Mycology Lab Center at Charles Nicole Hospital, Tunisia.** ## Technical Report A detailed technical report outlining the methodologies and findings is available on my ResearchGate profile: **Detection of Malaria from Microscopic Images**. ## 🚀 Streamlit App Concept ### 🎬 App Demo GIF .gif) ### 🔬 Bringing AI to Malaria Diagnosis This app is designed to **assist doctors** in diagnosing malaria from blood smear images using an interactive **Streamlit** interface. Developed with **custom ML tools** tailored for **Tunisian hospitals**, it adapts to real-world **microscopic imaging conditions** for accurate and reliable parasite detection. ### ⚡ How It Works - **Upload a Blood Smear Image:** Drag and drop your sample image into the app. - **AI-Powered Parasite Detection:** The model highlights malaria parasites with bounding boxes. - **Smart Classification:** Each detected parasite is labeled by type. - **Dynamic Visualization:** See instant stats, charts, and a clear infection diagnosis. ✅ **Intuitive. Fast. Designed for medical professionals.** ## Data Collection The dataset was gathered at the **Parasitology & Mycology Lab Center** of Charles Nicole Hospital using blood smears infected with malaria. In close collaboration with expert biologists and doctors over multiple sessions, we documented the data collection process. Images were captured with a **Canon EOS 1000D** and a **Zeiss Axiolab microscope** at **100x magnification**. Dataset Capture Setup (photo captured during sessions) Dataset Sample (photo captured during sessions) ## Upcoming README Updates **Note:** While the current technical report focuses on the core functionalities, additional f …