Optical Skin-fat Thickness Measurement Using Miniaturized Chip LEDs: A Preliminary Human Study

Abstract

We tested the feasibility of measuring fat thickness rising a miniaturized chip LED sensor module, testing 12 healthy female subjects. The module consisted of a single detector and four sources at four different source-detector distances (SD). A segmental curve-fitting procedure was applied, using ail empirical algorithm obtained by Monte-Carlo simulation, and fat thicknesses were estimated. These thicknesses were compared to computed-tomography (CT) results; the correlation coefficient between CT and optical measurements was 0.932 for bicep sites. The mean percentage error between the two measurements was 13.12%. We conclude that fat, thickness can be efficiently measured using the simple sensor module.