Chewing monitoring
Chewing is a core component of the intake process and micro-structure. Our early work focused on wearable/implant devices worn at/in the ear canal to capture food breakdown sounds conducted from mandible to skull. We found that the sounds are characteristic for certain foods and food types due to their material textures. For example, carrots and lettuce have wet-crisp cell structures that create specific acoustic patterns when chewed. Pattern classification of up to 19 foods was shown and is until today one of the highest number of foods discriminated by computers.
Chewing could help to estimate food weight and amount too, as chewing cycles could be monitored with the ear-worn sensor. Weight estimation error was ~20% for several foods, well below the confirmed errors of manual conducted self-reports. 
Currently, our work focused on chewing analysis integrated into smart eyeglasses. Results show that Electromyography (EMG) can be used to measure Temporalis muscle contraction during chewing and thus identify chewing in highly comfortable and unobtrusive, personalised 3D-printed eyeglasses.
Related projects
- WISEGlass
- Digital development & production
Publications
[publication-list tags=”chewing monitoring” separator=0 ]