Phonocardiography-based mitral valve prolapse detection with using fractional fourier transform

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Mitral Valve Prolapse (MVP) is a common cardiac condition, often asymptomatic but can lead to severe complications like heart failure.
• Early diagnosis of MVP is crucial for managing potential adverse outcomes.
• Phonocardiogram (PCG) signals offer valuable insights into heart valve function, making them suitable for MVP detection.

Purpose of the Study:

• To develop and evaluate a novel method for detecting Mitral Valve Prolapse (MVP) using phonocardiogram signals.
• To assess the efficacy of signal processing techniques and machine learning classifiers in identifying MVP.

Main Methods:

• Phonocardiogram signals were denoised and segmented.
• Feature extraction was performed using the Fractional Fourier Transform (FrFT).
• Features were refined using a Moving Logarithmic Median Window (MLMW) and selected based on distance criteria. Classification was conducted using K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) algorithms.

Main Results:

• The Support Vector Machine (SVM) classifier demonstrated superior performance.
• Achieved high accuracy (96.25% ± 2.43%), sensitivity (98.5% ± 3.37%), specificity (94.0% ± 5.16%), precision (96.0% ± 3.44%), kappa (92.5% ± 4.86%), and f-score (96.6% ± 2.34%).
• The A-test method was employed for experimental validation on a database of 15 prolapsed and 6 non-prolapsed subjects.

Conclusions:

• The proposed method effectively detects Mitral Valve Prolapse (MVP) using phonocardiogram signals.
• The combination of FrFT feature extraction and SVM classification provides a robust approach for MVP diagnosis.
• This technique holds potential for improving early detection and management of MVP complications.