Related Experiment Video
Updated: Jul 18, 2026
06:07
Analyzing Long-Term Electrocardiography Recordings to Detect Arrhythmias in Mice
Published on: May 23, 2021
Phonocardiographic signal analysis method using a modified hidden Markov model
Ping Wang1, Chu Sing Lim, Sunita Chauhan
1Biomedical Engineering Research Centre, Nanyang Technological University, 50 Nanyang Drive, Research Techno Plaza, 6th Storey, XFrontiers Block, Singapore 637553, Singapore.
Annals of Biomedical Engineering
|December 16, 2006
Summary
This study introduces a novel system using Mel-frequency cepstral coefficients (MFCC) and hidden Markov models (HMM) for accurate heart sound classification. The developed method significantly improves diagnostic interpretation for cardiovascular analysis.
Area of Science:
- Cardiology
- Biomedical Engineering
- Signal Processing
Background:
- Auscultation is a key diagnostic tool for cardiovascular assessment.
- Accurate heart sound classification is crucial for auscultative diagnosis.
- Existing methods require efficient feature extraction and classification techniques.
Purpose of the Study:
- To develop and evaluate a system for interpreting heart sounds using pattern recognition.
- To compare the efficacy of different feature extraction methods for heart sound analysis.
- To assess the performance of Mel-frequency cepstral coefficients (MFCC) combined with hidden Markov models (HMM) for heart sound classification.
Main Methods:
- Heart sound cycles were pre-processed for feature extraction.
- Feature extraction was performed using time-domain features, short-time Fourier transforms (STFT), and MFCC.
- A hidden Markov model (HMM) was employed for automatic classification of heart sounds.
- The system was evaluated on 1398 datasets from 41 subjects.
Main Results:
- Mel-frequency cepstral coefficients (MFCC) demonstrated superior performance in feature extraction compared to other methods.
- The MFCC-HMM system achieved high classification accuracy: sensitivity >= 0.952 and specificity >= 0.953.
- The system successfully classified normal heart sounds and various murmur characteristics.
- Constituent characteristics of heart sounds were effectively evaluated.
Conclusions:
- The proposed MFCC-HMM system offers improved interpretative information for heart sound analysis.
- The system shows potential as an objective tool to aid clinicians in cardiovascular diagnosis.
- This approach enhances the diagnostic capabilities of phonocardiography through advanced pattern recognition.
