Publikation

Monitoring Dietary Behavior with a Smart Dining Tray

Bo Zhou, Jingyuan Cheng, Paul Lukowicz, Attila Reiss, Oliver Amft

In: IEEE Pervasive Computing 14 Seiten 46-56 IEEE 2015.

Abstrakt

In this article, the authors investigate a new source of information for dietary monitoring: pressure distribution on the surface underneath dining plates. Pressure sensing has been used to consider the weight of the eaten food. The core idea behind their work is that dynamic pressure information can also be used to distinguish between various cutlery-related activities, such as cutting, poking, stirring, or scooping. The authors show how to spot such individual actions in continuous datastreams, assign them to specific containers (main plate, salad bowl), count them (how many bites taken), and relate them to different abstract food categories. They consider two sensing modalities: (1) textile pressure-sensor matrix technology facilitating a "smart tablecloth" that looks and feels like a standard tablecloth but provides detailed information on the spatial and temporal pressure; and (2) standard force sensitive resistor (FSR) sensors placed underneath a rigid tray. They also present the results of a new, comprehensive study with 10 subjects, each of whom consumed a total of eight meals chosen from 17 possible main dishes with six possible side dishes; results show an average accuracy of up to 94 percent. This article is part of a special issue on pervasive food.

Weitere Links

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence