Dietary Monitoring for Diabetes & Obesity: Detection and Quantification of Chewing Behavior - Background and Phenomenology Technical Report MPCL-08-07

An important component of maintaining a healthy lifestyle is a healthy diet. Dietary monitoring can be effected using wearable sensors or computers, networked to achieve dense communication with one or more monitoring and analysis nodes. Dietary monitoring sensor modalities include motion, acoustic signals, electrical signals from muscles, and analysis of facial imagery from video sequences. In this technical summary, we overview the role of video-based detection and quantification of chewing behaviors for dietary monitoring. Section 1 overviews the context of mastication in health diet, and details techniques by which chewing has been detected in previous work. Section 2 discusses the neurophysiological basis for chewing, and overview the work of Cutler et al. in the detection and quantification of periodic motion. Original papers are abstracted, and bibliographies are reproduced at the end of this summary. Section 3 overviews the application of Cutler’s work to the analysis of video imagery containing facial regions in which chewing behavior is manifest. Preliminary conclusions are presented in Section 4.