KMID : 1144120140040040378
|
|
Biomedical Engineering Letters 2014 Volume.4 No. 4 p.378 ~ p.387
|
|
Wearable sensing for liquid intake monitoring via apnea detection in breathing signals
|
|
Dong Bo
Biswas Subir
|
|
Abstract
|
|
|
Purpose: Appropriate amount of liquid intake is crucial for maintaining human physiological operations. Traditionally, researchers have used self-reported questionnaires for estimating daily liquid intake, which has been proven to be unreliable. In this study, we developed an instrumented system for liquid intake monitoring to reduce estimation subjectivity by complementing self-reporting information with instrumented data.
Methods: Liquid intake can be detected by the way of detecting a person¡¯s swallow events. The system works based on a key observation that a person¡¯s otherwise continuous breathing process is interrupted by a short apnea when a swallow occurs as a part of the intake process. We detect the swallows via recognizing apneas extracted from breathing signal captured by a wearable sensor chest-belt. Such apnea detection is performed using matched filters and machine learning mechanisms with both time and frequency domain features. Spectrum analysis, artifact handling, and iterative template refinement were also proposed, analyzed and experimented with.
Results: It is demonstrated that the proposed matched filter method on an average can provide true positive rates up to 82.81% and false positive rates as low as 3.31%. It is also demonstrated that the machine learning method using Decision Tree (J48) provides the best true positive rates up to 97.5% and false positive rates as low as 0.7%.
Conclusions: The experiments and analysis suggest that the proposed liquid intake monitoring system and algorithms through breathing signal shows potential for being used for liquid intake monitoring.
|
|
KEYWORD
|
|
Wearable sensing, Swallow detection, Liquid intake monitoring, Breathing pattern analysis, Matched filter
|
|
FullTexts / Linksout information
|
|
|
|
Listed journal information
|
|
|