Umvr873/ML-Based-Support-System-for-Early-Malaria-Diagnosis

This system uses a MobileNetV2 transfer learning model to classify microscopic blood smear images as either Parasitized or Uninfected. The goal is to provide real-time diagnostic support for early malaria screening, especially in low-resource settings. # ML-Based Support System for Early Malaria Diagnosis This is a Streamlit web application for malaria detection using microscopic blood smear images. ## Required model files Place these files inside the `model/` folder: - `best_malaria_mobilenetv2.keras` - `class_mapping.json` ## Run locally ```bash pip install -r requirements.txt streamlit run app.py ``` ## Class Mapping The trained model used this mapping: ```python {'Parasitized': 0, 'Uninfected': 1} ``` Therefore: - Prediction probability = 0.5 means Uninfected ## Disclaimer This application is a diagnostic support system only. It is not a replacement for professional medical diagnosis.