AI Pneumonia Detector

Project description

AI Pneumonia Detector compares two machine learning approaches for pneumonia detection from radiographic images: CNN (TensorFlow/PyTorch) and KNN (Scikit-learn). The user can test X-rays (examples provided) and get an instant prediction, along with detailed metrics (precision, recall, F1-score), confusion matrices, and ROC curves to understand model performance. The Vue.js/TailwindCSS front-end ensures a clear interface, while the Python/FastAPI (Uvicorn) back-end orchestrates pre-processing and inference. The project highlights medical AI pedagogy, reproducibility (Jupyter, nbconvert/nbformat), and optimization (Pandas/NumPy, memory management psutil/GPUtil). A concrete demonstration of data science applied to health with open-source technologies.