Electro Well: Fat and Hydration Detection using NIR Sensor

This paper shows a simple, inexpensive approach of estimating hydration and fat percentage using Near-Infrared (NIR) light. Determination of hydration and fat level is important in medical diagnostics and personal health assessment. Laboratory analysis or large, expensive equipment such as BIA machines and DEXA scanners are very successful methods, though not practical for daily use or fast monitoring. Our method applies the small, noninvasive setup with an AS7263 six-channel NIR sensor and an Arduino. It measures how human tissue reflects light from 850 nm to 1020 nm. Different wavelengths give different interactions in water and fat; thus, they deliver important information about body composition. We designed a simple calculation model that converts the raw data of reflection into hydration and fat percentages using normalized reflectance and the water-to-fat ratio. These experiments have clearly established a relationship between absorbed wavelengths and the composition of the body, suggesting that the NIR optical method can be used in portable health monitoring in realtime. This paper describes the hardware, algorithm, calibration of the system, results, and limits for a comprehensive budget-friendly biomedical sensing approach within the remit of researchers, students, and wearable health applications.