Towards the Development of a Mobile Phonopneumogram: Automatic Breath-Phase Classification Using Smartphones

Area of Science:

• Biomedical Engineering
• Respiratory Physiology
• Medical Technology

Background:

• Accurate labeling of breath phases is crucial for analyzing respiratory sounds.
• Traditional methods rely on airflow or volume signals, which are not always accessible.
• There is growing interest in developing smartphone-based systems for respiratory sound analysis.

Purpose of the Study:

• To propose and evaluate an optical method using a smartphone camera to capture chest movements for breath phase classification.
• To assess the correlation between smartphone-derived chest movement signals and spirometer-measured volume.
• To determine the accuracy of smartphone-based breath phase classification, especially during complex breathing patterns.

Main Methods:

• Acquired simultaneous spirometer and smartphone-based chest movement signals from 13 healthy volunteers.
• Utilized a smartphone camera to record chest movements during various breathing frequencies and airflow levels.
• Applied simple linear regression to the chest movement signal to label breath phases based on signal slope.

Main Results:

• The smartphone-acquired chest movement signal showed a high correlation with reference spirometer volume (ρ = 0.960 ± 0.025).
• The proposed method achieved 100% accuracy in classifying breath phases.
• The classification scheme effectively handled challenging breathing patterns, including non-breath events like swallowing, talking, and coughing, as well as irregular breathing.

Conclusions:

• A smartphone camera can provide a reliable chest movement signal for breath phase classification in respiratory sound analysis.
• This optical approach offers a feasible, portable, and inexpensive solution for developing phonopneumograms.
• The system demonstrates potential for accurate respiratory sound analysis in diverse and complex breathing conditions.