Towards a Non-intrusive Self-management System for Asthma Control Using Smartphones
A noise-robust algorithm for segmentation of breath events during continuous speech is presented. The built-in microphone of a smartphone is used to capture the speech signal (voiced and breath frames) under conditions of a relatively noisy background. A template matching approach, using mel-cepstrograms, is adopted for constructing several similarity measurements to distinguish between breath and non-breath frames. Breath events will be used for lung function regression.
Keywordsasthma control breath detection MFCC mel-cepstrogram lung function
Unable to display preview. Download preview PDF.
- 1.The global initiative for asthma (GINA), http://www.ginasthma.org/
- 2.Bravo, J.: m-Health: Mobile Computing and Health Monitoring. J. of Computation in Biosciences and EngineeringGoogle Scholar
- 3.Larson, E.C., Goel, M., Boriello, G., Helthe, S., Rosenfeld, M., Patel, S.N.: SpiroSmart: Using a microphone to measure lung function on a mobile phone. In: UbiComp, pp. 280–289 (2012)Google Scholar
- 5.Hongzhi, W., Yuchao, X., Meijing, L.: Study on the MFCC similarity-based voice activity detection algorithm. In: AIMSEC, pp. 4391–4394 (2011)Google Scholar