Related Experiment Video
Updated: Apr 6, 2026
06:21
Concept Development and Use of an Automated Food Intake and Eating Behavior Assessment Method
Published on: February 19, 2021
6.5K
Audio-based detection and evaluation of eating behavior using the smartwatch platform.
Haik Kalantarian1, Majid Sarrafzadeh1
1Wireless Health Institute, Department of Computer Science, University of California, Los Angeles, United States.
Computers in Biology and Medicine
|August 5, 2015
Summary
Smartwatches can effectively identify eating and drinking behaviors using built-in microphones. This technology shows high potential for non-invasive nutrition monitoring, achieving 94.5% accuracy in classifying bites and swallows.
Area of Science:
- Biomedical Engineering
- Wearable Technology
- Human-Computer Interaction
Background:
- Smartwatches offer a user-accepted platform for health applications.
- Existing custom hardware for health monitoring faces lower user acceptance.
- The integration of sensors in smartwatches enables novel physiological monitoring capabilities.
Purpose of the Study:
- To develop and evaluate signal-processing techniques for identifying eating and drinking events using smartwatch microphones.
- To assess the feasibility of smartwatches for nutrition monitoring.
- To analyze the overall applicability of a smartwatch-based system for food intake monitoring.
Main Methods:
- Utilized the built-in microphone of a smartwatch to capture audio signals during eating and drinking.
- Developed signal-processing algorithms for the classification of chews, swallows, talking, and ambient noise.
- Conducted a survey to evaluate user acceptance and potential for nutrition monitoring via smartwatches.
- Collected 250 audio samples for system evaluation.
Main Results:
- Achieved high classification accuracy between different food bites (apple, potato chip) and water swallows.
- Successfully differentiated eating/drinking sounds from talking and ambient noise.
- Attained an F-measure of 94.5% for the classification task.
- Survey results indicated positive user perception of smartwatches for nutrition monitoring.
Conclusions:
- Smartwatch-based audio analysis is an effective method for detecting food intake events.
- The developed signal-processing techniques demonstrate high efficacy for chews and swallows identification.
- Smartwatches present a promising, user-friendly platform for unobtrusive nutrition monitoring.
