Non-Invasive Anemia Detection: A Multichannel PPG-Based Hemoglobin Estimation with Explainable Artificial Intelligence
A new AI system uses light sensors on the skin to estimate hemoglobin, eliminating the need for blood draws.
Deep Dive
Researchers Garima Sahu and Nachiket Tapas developed a non-invasive AI framework for anemia screening. It uses a four-wavelength photoplethysmography (PPG) sensor (660-940 nm) and a gradient boosting model to estimate hemoglobin concentration from skin signals. The system achieved a mean absolute error of 8.50 g/L and uses SHAP for explainability. This allows for preliminary, needle-free anemia screening and continuous monitoring using wearable sensor data.
Why It Matters
It enables large-scale, accessible health screening and continuous monitoring for a condition affecting billions globally.