Research and application of heart sound alignment and descriptor

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Cardiovascular disease (CVD) diagnosis relies heavily on auscultation.
• Traditional heart sound (HS) analysis often focuses on specific components or features.
• A need exists for holistic HS analysis methods to improve CVD diagnosis.

Purpose of the Study:

• To develop and validate a novel method for comparing the overall acoustical characteristics of heart sounds.
• To assess the efficacy of different alignment schemes, specifically binary coding and N-gram, for HS analysis.
• To evaluate the potential of this approach for simplifying non-invasive cardiovascular disease diagnosis.

Main Methods:

• Heart sounds (HS) are treated as a whole, rather than focusing on individual components.
• Acoustical characteristics of HS are compared using sequence alignment and similarity index evaluation.
• Technologies adapted include musical instrument digital interface (MIDI), binary coding, N-gram, Lempel-Ziv (L-Z) complexity, and super-symmetric comparison distance (SCD).

Main Results:

• Thorough investigation of binary and N-gram alignment schemes for heart sound analysis.
• Testing results demonstrate the superiority of the N-gram alignment approach.
• The proposed method successfully compares the similarity of acoustical characteristics between heart sounds.

Conclusions:

• The N-gram based heart sound analysis method offers a novel and effective approach for cardiovascular disease diagnosis.
• This method simplifies the diagnostic process, moving beyond traditional auscultation.
• The approach holds promise for widespread application in e-home healthcare due to its non-invasive nature.