Gert Mertes1, Tom Croonenborghs, Bart Vanrumste1, Hans Hallez1
13:25 - 14:15 | Thu 16 Feb | Ballroom D | ThRPF.12
In this work, we propose the use of a glasses mounted accelerometer to detect chewing motion in the elderly. Data from 13 elderly was collected during their daily meals. This data is used to evaluate a k-Nearest Neighbor classifier. An average precision of 61.4 and average recall of 56 is achieved on the entire dataset of 13 people, while an average precision and recall of 74.6 and 62.1 respectively is achieved on a subset of 9 out of 13 people.