Compressed Sensing Data with Performing Audio Signal Reconstruction for the Intelligent Classification of Chronic...

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Respiratory Medicine

Background:

• Chronic Obstructive Pulmonary Disease (COPD) significantly impacts global health, necessitating improved diagnostic and monitoring tools.
• Early detection of lung function decline and exacerbation prediction in COPD patients remain critical challenges.
• Scalable, data-driven AI methods are crucial for advancing modern COPD healthcare.

Purpose of the Study:

• To establish experimental foundations for AI-driven diagnosis and monitoring of COPD using biomedical data.
• To investigate multi-resolution analysis, compression, and machine classification of lung audio signals for COPD detection.
• To develop next-generation diagnostic systems for future COPD healthcare.

Main Methods:

• Acquisition and generation of biomedical observation data for signal analysis and machine learning.
• Multi-resolution analysis and compression of lung audio signals.
• Machine classification of audio signals into 'Healthy' vs. 'COPD' and 'Healthy', 'COPD', vs. 'Pneumonia' categories.
• Signal reconstruction to ensure data integrity.

Main Results:

• High accuracy achieved in classifying 'Healthy' or 'COPD' conditions using machine learning classifiers.
• Promising accuracy demonstrated in classifying 'Healthy', 'COPD', or 'Pneumonia' conditions.
• Selected machine learning models exhibited strong performance across diverse metrics.

Conclusions:

• The study presents a promising AI-based approach for the intelligent diagnosis and monitoring of COPD using lung audio signals.
• The developed methods show potential for accurate classification of respiratory conditions, including COPD and pneumonia.
• Future work will integrate multi-modal sensing and data fusion for enhanced diagnostic capabilities.