Related Experiment Video
Updated: Oct 10, 2025
04:04
Asthma Detection Research Based on Voice Signal Processing and Machine Learning
Published on: July 22, 2025
521
COVID-19 Biomarkers in Speech: On Source and Filter Components
Summary
This study identifies biomarkers in human speech for COVID-19 detection. Analyzing cough and breathing sounds with source-filter separation shows promise for identifying positive cases.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Infectious Disease Diagnostics
Background:
- The COVID-19 pandemic necessitates novel diagnostic tools.
- Human speech signals contain physiological information potentially indicative of respiratory infections.
Purpose of the Study:
- To identify biomarkers for COVID-19 detection within human speech signals.
- To analyze source-filter components of speech, cough, and breathing sounds for COVID-19 indicators.
Main Methods:
- Source-filter separation techniques (cepstral, phase domain) were applied to speech signals.
- Neural networks were utilized to detect COVID-19 positive subjects based on separated signal components.
- Vowels, cough, and breathing sounds were comparatively analyzed.
Main Results:
- Source-filter separation effectively distinguished between healthy and COVID-19 positive subjects.
- Specific vocal tract and excitation biomarkers were identified in speech signals of COVID-19 patients.
- Cough and breathing sounds provided additional diagnostic information.
Conclusions:
- Human speech analysis, particularly cough and breathing sounds, offers a viable non-invasive method for COVID-19 detection.
- Source-filter separation techniques are effective for extracting relevant biomarkers from speech.
- Further research can refine these methods for clinical application.
