Wearable diet monitoring through breathing signal analysis

Abstract

This paper presents the design, system structure and performance for a wireless and wearable diet monitoring system. Food and drink intake can be detected by the way of detecting a person's swallow events. The system works based on the key observation that a person's otherwise continuous breathing process is interrupted by a short apnea when she or he swallows as a part of solid or liquid intake process. We detect the swallows through the difference between normal breathing cycle and breathing cycle with swallows using a wearable chest-belt. Three popular machine learning algorithms have been applied on extracted time and frequency domain features. It is shown that high detection performance can be achieved with only few features.

Wearable Food Intake Monitoring Technologies: A Comprehensive Review

“…(a) Acoustic approach: sensor prototype with ear pad cushion (left) with the attachment of the foam cushion to the sensor (right) (adapted from [49]); (b) Visual approach: eButton system with a camera embedded in a button (left) and the smart phone (right) (adaped from [50]); (c) Inertial approach: MEMS gyroscope prototype on the wristband for wrist motion tracking. This self-contained system includes a microprocessor, battery, gyroscope, LCD, memory and USB port connection (adapted from [51]); (d) EMG/EGG-based approach: EGG sensors are attached to a neoprene collar (left) and the collar is fastened to the neck of the subject by the Velcro (right) (adapted from [52]); (e) Piezoelectric approach: piezoelectric film-embedded necklace to detect muscular contractions during swallowing (adapted from [53]); (f) Fusion approach: an example of the vision-acoustic approach -the microphone is embedded in the housing for detecting chewing/eating events and the camera faces toward the food plate for food detection and classification (adapted from [54]); (g) Electrical proximity sensing: the prototype consists of a textile neckband with 4 capacitive sensors for the detection of single swallows [55]; (h) Respiratory inductance plethysmography: the prototype is worn on the chest to collect the breathing signal and and to transfer it to a smart phone through Bluetooth (adapted from [56]). …”

“…It represents the change in cross-sectional area of the body when breathing. Dong et al [136] and Dong and Biswas [56,137] constructed an RIP belt shown in Figure 4h containing a breathing signal sensing system that embeds swallowing signatures in this respiratory signal for swallowing detection. The experimental procedure for this RIP belt consists of only six subjects and is limited to the laboratory condition, and its evaluation result has a moderate accuracy of 88% to 73.33% with an SVM classifier; however, this approach has the advantage of no contact with skin and shows an exceptional novelty in wearable food intake monitoring.…”

Mobile Ecological Momentary Diet Assessment Methods for Behavioral Research: Systematic Review (Preprint)

“…image-assisted and image-based assessment [14][15][16][17][18][19][20][21][22][23][24] and the detection of food intake by biomechanical sensors or hand-held devices [25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42]. Significant progress has been made in image-assisted and image-based food recording that has resulted in the improved accuracy of dietary self-report [15,24].…”

“…Significant progress has been made in image-assisted and image-based food recording that has resulted in the improved accuracy of dietary self-report [15,24]. Similarly, wearable technology including gyroscopes, microphones, and mechanical or electrical impedance sensors have been adapted to detect wrist or hand motion [25][26][27][28][29][30][31] or patterns of chewing or swallowing indicative of food intake (eg, number of bites) [32][33][34][35][36][37][38][39][40][41]. However, design and proof-of-concept data suggest the automation of dietary assessment remains out of reach for the time being.…”

“…However, design and proof-of-concept data suggest the automation of dietary assessment remains out of reach for the time being. For instance, the mean detection accuracy of image detection and wearable devices have been acceptable in controlled settings (range = 73-99%) [18,26,30,31,34,35,[38][39][40][41][42][43][44][45][46][47][48] but limited testing has been done in natural settings [27].…”