Design and evaluation of a parametric model for cardiac sounds

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Heart sound analysis is crucial for diagnosing cardiovascular diseases.
• Phonocardiograms (PCG) are essential for auscultative diagnosis.
• Accurate PCG modeling aids in disease detection.

Purpose of the Study:

• To propose a novel parametric heart sound model.
• To accurately represent normal and pathological cardiac audio signals (PCG).
• To rigorously evaluate the proposed model's performance.

Main Methods:

• PCG signal modeled as deterministic and stochastic components.
• Deterministic part modeled using Matching Pursuit (MP) algorithm.
• Stochastic (residual) part modeled using Linear Predictive Coding (LPC).

Main Results:

• Objective evaluation showed ~5% Root-Mean-Square Difference.
• Subjective evaluation with medical experts confirmed high accuracy.
• The model provides a highly accurate approximation of real heart sounds.

Conclusions:

• The proposed parametric heart sound model accurately represents PCG signals.
• The model demonstrates high performance in both objective and subjective evaluations.
• This represents a significant advancement in heart sound modeling and analysis.