Smooth bandpass empirical mode decomposition with rolling ball sifting for extracting carotid bruits and heart sounds

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiovascular Acoustics

Background:

• Carotid bruits are audible indicators of turbulent blood flow in atherosclerotic stenosis.
• Extracting both bruits and heart sounds with high fidelity is challenging for computer-aided analysis.
• Existing methods struggle with the complex, intermittent nature of these sounds.

Purpose of the Study:

• To develop a novel signal processing method for high-fidelity extraction of carotid bruits and heart sounds.
• To address the nontrivial problem of separating superimposed acoustic signals in the time domain.
• To enable more accurate computer-aided analysis of cardiovascular and cerebrovascular conditions.

Main Methods:

• Proposed a smooth bandpass empirical mode decomposition (EMD) method operating in the time domain.
• Utilized a rolling ball algorithm to sift local extrema at chosen time-scales for bandpass filtering.
• Employed monotone piecewise cubic spline interpolation to smooth local zeros of the signal components.

Main Results:

• The smooth bandpass EMD method successfully extracted both carotid bruits and heart sounds.
• The extracted components were visually smooth oscillating signals, indicating high fidelity.
• Preliminary results demonstrate the method's efficacy in separating complex acoustic signals.

Conclusions:

• The proposed smooth bandpass EMD method offers an effective solution for extracting carotid bruits and heart sounds.
• This technique enhances the potential for accurate computer-aided analysis of atherosclerotic stenosis.
• The method provides a foundation for improved diagnostic tools in cardiovascular and cerebrovascular medicine.