Artificial intelligence and machine learning for objective intake monitoring
More details
Hide details
Department of Informatics and Telematics, Harokopio University of Athens, Athens, Greece
Department of Electronics and Computer Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece
Publication date: 2022-05-27
Public Health Toxicol 2022;2(Supplement 1):A16
The progress in artificial intelligence and machine learning during the last decade has enabled the development of new and improved intake monitoring methods. Specifically, sensor-based objective monitoring has the potential to provide more accurate intake indicators compared to questionnaires, but can also provide detailed in-meal intake parameters that today are available only through time consuming, video-based manual annotations.

This presentation aims at providing an overview of our work on objective intake monitoring.

We present two types of methods, one based on commercially-available smartwatches and one based on a prototype ear-worn chewing sensor. Smartwatch-based methods take advantage of the inertial sensors (triaxial accelerometer and gyroscope) available in modern smartwatches to detect intake gestures. It is shown that although individual movements are not sufficient to detect intake, modelling sequences of movements leads to accurate detection. For the chewing sensor, we evaluate an in-ear microphone as well as a photoplethysmography sensor for chewing detection.

We present results on several annotated datasets. The smartwatch-based intake monitoring methods achieve an accuracy of F1 score of 0.92 for detection of intake cycles and 0.96 for meals. Chewing sensor methods achieve 0.96 weighted accuracy and 0.91 F1 score for meal detection.

Kyritsis K, Diou C, Delopoulos A. A Data Driven End-to-End Approach for In-the-Wild Monitoring of Eating Behavior Using Smartwatches. IEEE J Biomed Health Inform. 2021;25(1):22-34. doi:10.1109/JBHI.2020.2984907
Kyritsis K, Diou C, Delopoulos A. Modeling Wrist Micromovements to Measure In-Meal Eating Behavior From Inertial Sensor Data. IEEE J Biomed Health Inform. 2019;23(6):2325-2334. doi:10.1109/JBHI.2019.2892011
Papapanagiotou V, Diou C, Zhou L, van den Boer J, Mars M, Delopoulos A. A Novel Chewing Detection System Based on PPG, Audio, and Accelerometry. IEEE J Biomed Health Inform. 2017;21(3):607-618. doi:10.1109/JBHI.2016.2625271
Papapanagiotou V, Diou C, Delopoulos A. Self-Supervised Feature Learning of 1D Convolutional Neural Networks with Contrastive Loss for Eating Detection Using an In-Ear Microphone. Annu Int Conf IEEE Eng Med Biol Soc. 2021;2021:7186-7189. doi:10.1109/EMBC46164.2021.9630399