Non_invasive-detection-of-anemia

Graduation project

Overview

This project implements a two-phase diagnostic system for blood analysis:

Phase 1: Binary anemia detection using conjunctiva, palm, and nail images

Phase 2: Hemoglobin (Hgb) level regression using late fusion technique

Achieved state-of-the-art results with:

Phase 1: 99.5% accuracy (palm), 99.4% (nails), 99.38% (conjunctiva)

Phase 2: 93% R² score and 0.3 MSE for Hgb level prediction