17 May 2018 Eating habits characterization with NIR spectroscopy and bioimpedance wearable sensor
Author Affiliations +
Abstract
We propose multimodal sensor and algorithm for automatic recognition of a food intake based on glycemic response. Embedding this sensor in a wearable device makes it possible to count number of meals at a given time and to generate personalized statistical pattern of eating habits. This pattern may have significant impact on both personal health care and big-data-driven social engineering. We use near-infrared diffuse reflectance spectroscopy, bioimpedance measurements, and binary classification for non-invasive continuous glucose trend measurements and Fourier transform based time frequency analysis of glycose trends for characterization of eating patterns and prediction of digestive system abnormalities. We tested the sensor in a series of experiments with the certain type of food and achieved 45% average accuracy of a food intake recognition with the random noise level being at 25%.
© (2018) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Vladislav Lychagov, Konstantin Pavlov, Mikhail Popov, "Eating habits characterization with NIR spectroscopy and bioimpedance wearable sensor", Proc. SPIE 10685, Biophotonics: Photonic Solutions for Better Health Care VI, 106852V (17 May 2018); doi: 10.1117/12.2301019; https://doi.org/10.1117/12.2301019
PROCEEDINGS
6 PAGES


SHARE
Back to Top