Research Article
Features of Audio Frequency Content of Respiration to Distinguish Inhalation from Exhalation
@INPROCEEDINGS{10.1007/978-3-031-43135-7_14, author={Souhail Katti and Federica Aveta and Saurav Basnet and Douglas E. Dow}, title={Features of Audio Frequency Content of Respiration to Distinguish Inhalation from Exhalation}, proceedings={Bio-inspired Information and Communications Technologies. 14th EAI International Conference, BICT 2023, Okinawa, Japan, April 11-12, 2023, Proceedings}, proceedings_a={BICT}, year={2023}, month={9}, keywords={Fourier FFT LabView MATLAB eupneic breathing chronic monitoring}, doi={10.1007/978-3-031-43135-7_14} }- Souhail Katti
Federica Aveta
Saurav Basnet
Douglas E. Dow
Year: 2023
Features of Audio Frequency Content of Respiration to Distinguish Inhalation from Exhalation
BICT
Springer
DOI: 10.1007/978-3-031-43135-7_14
Abstract
The life-sustaining function of respiration becomes impaired by diseases that occur more with old. A system that monitors the inhalations and exhalations of the respiratory cycle could raise an alert when abnormal patterns or prolonged disruptions are detected. Noninvasive methods are suitable to chronically monitor respiration. Methods include analyzing audio sounds generated during respirations and analyzing changes in the volume of the thorax or abdomen. In casual observations of eupneic breathing, inhalation often sounds different from exhalation, though may be quite similar. One of the challenges for signal processing is to distinguish inhalation from exhalation based on only the audio. The purpose of this study was to find a method of analyzing the audio frequency content that could differentiate the inhalation and exhalation. Volunteer subjects were recruited to record audio during eupneic respiration for analysis. To classify the timing of each inhalation and exhalation, both respiratory sounds and volume changes of the thorax were simultaneously recorded. The audio files were analyzed by Fast Fourier Transform (FFT) to determine the frequency content. Features of the frequency power spectrum were found that appear promising for distinguishing inhalation and exhalation. Such differences could be used to characterize audio respiratory signals and improve the monitoring of individuals at risk for impaired respiratory function.
