A robust heart sound segmentation algorithm for commonly occurring heart valve diseases

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Accurate segmentation of heart sounds (phonocardiogram) is crucial for diagnosing valvular heart diseases.
• Existing methods often rely on electrocardiographic (ECG) signals, necessitating complex instrumentation.
• There is a need for simpler, more efficient segmentation techniques.

Purpose of the Study:

• To develop an automatic heart sound segmentation algorithm that does not require an auxiliary ECG signal.
• To improve the accuracy and reduce the computational complexity of heart sound segmentation.
• To evaluate the algorithm's performance across various conditions and pathological cases.

Main Methods:

• Developed an automatic segmentation algorithm utilizing biomedical domain features.
• The algorithm processes phonocardiogram signals without requiring simultaneous ECG recordings.
• Performance was evaluated using nine pathological heart sound cases and normal heart sounds.

Main Results:

• Achieved an overall accuracy of 97.47% in heart sound segmentation.
• Demonstrated superior performance compared to two competing techniques.
• Showcased robustness against additive white Gaussian noise at various signal-to-noise ratio (SNR) levels.

Conclusions:

• The proposed automatic segmentation method offers a more accurate and computationally efficient alternative for heart sound analysis.
• Eliminating the need for ECG signals simplifies instrumentation and broadens applicability.
• The algorithm shows promise for reliable detection of valvular heart diseases.