Contributed Paper (Oral)
09:35 - 11:05 | Mon 5 Mar | Treasure Island ABC | MoAT1
Obesity is one of the leading cause of a set of chronic diseases. Successful weight-loss interventions depend on changing the unhealthy lifestyle and maintaining awareness of individual's eating habits. Recent nutritional behavior management systems are considered as open loop systems. In this study, we propose a closed loop strategy through monitoring and evaluation of various food intake activities. Wireless surface electromyogram (sEMG) was deployed in order to differentiate between meal, snack and drink activities through a multistage classification system. The proposed algorithm was able to discriminate between eating and drinking activities with accuracy of 96% using time domain features and K-Nearest Neighbour classifier (KNN). Furthermore using Hilbert transform based classifier, we scored 97.5% accuracy for drinking/saliva swallowing classification and 93% accuracy for swallowing/chewing activities classification. These results suggest high efficiency of the proposed methodology in identifying the ingestive behaviour.
No information added