Fat Sensing Using Low-Cost Near-Infrared Spectroscopy

Abstract

Cardiovascular disease (CVD) is the leading cause of death globally and is a major financial burden on healthcare systems. Dietary management in the areas of fat and cholesterol is seen as a dominant risk reduction strategy. It is estimated that 80% of CVD may be preventable. Cutting down on dietary fat reduces the risk of CVD. This work focuses on exploring the extent to which the fat content of different foodstuffs can be determined using a low-cost Near-Infrared (NIR) Spectrometer. Preprocessing algorithms to manipulate data have been developed, classification models of materials' chemical makeup have been constructed using a Random Forest algorithm, and vegetable fat spreads which vary in levels of fat and cholesterol have been distinguished with an accuracy of over 99%.