Wheeze detection based on time-frequency analysis of breath sounds

Area of Science:

• Respiratory Medicine
• Biomedical Engineering
• Signal Processing

Background:

• Abnormal breath sounds, such as wheezes, are key indicators of obstructive pulmonary diseases.
• Accurate detection and monitoring of wheezes are crucial for patient management.

Purpose of the Study:

• To develop an automated technique for wheeze detection and monitoring.
• To utilize spectral analysis for identifying wheeze characteristics.

Main Methods:

• Recorded wheezes from 13 patients diagnosed with asthma, COPD, and pneumonia.
• Constructed a time-frequency wheeze detector (TF-WD) based on spectral features.
• Evaluated TF-WD performance against expert clinical auscultation and artificial noise.

Main Results:

• The TF-WD demonstrated efficient performance in detecting wheezes.
• The system exhibited high robustness against artificial noise.
• Comparison with expert auscultation validated the detector's accuracy.

Conclusions:

• The developed TF-WD provides an effective automated solution for wheeze detection.
• This technology holds potential for improved monitoring of obstructive pulmonary diseases.
• The system's noise robustness suggests clinical applicability in diverse environments.