Limits...
A framework for automatic heart sound analysis without segmentation.

Yuenyong S, Nishihara A, Kongprawechnon W, Tungpimolrut K - Biomed Eng Online (2011)

Bottom Line: Geometric mean was used as performance index.Average classification performance using ten-fold cross-validation was 0.92 for noise free case, 0.90 under white noise with 10 dB signal-to-noise ratio (SNR), and 0.90 under impulse noise up to 0.3 s duration.Further work include building a new training set recorded from actual patients, then further evaluate the method based on this new training set.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Communication and Integrated Systems, Tokyo Institute of Technology, Japan 2-12-1-W9-108 Ookayama, Meguro-ku, Tokyo, Japan. toey123@gmail.com

ABSTRACT

Background: A new framework for heart sound analysis is proposed. One of the most difficult processes in heart sound analysis is segmentation, due to interference form murmurs.

Method: Equal number of cardiac cycles were extracted from heart sounds with different heart rates using information from envelopes of autocorrelation functions without the need to label individual fundamental heart sounds (FHS). The complete method consists of envelope detection, calculation of cardiac cycle lengths using auto-correlation of envelope signals, features extraction using discrete wavelet transform, principal component analysis, and classification using neural network bagging predictors.

Result: The proposed method was tested on a set of heart sounds obtained from several on-line databases and recorded with an electronic stethoscope. Geometric mean was used as performance index. Average classification performance using ten-fold cross-validation was 0.92 for noise free case, 0.90 under white noise with 10 dB signal-to-noise ratio (SNR), and 0.90 under impulse noise up to 0.3 s duration.

Conclusion: The proposed method showed promising results and high noise robustness to a wide range of heart sounds. However, more tests are needed to address any bias that may have been introduced by different sources of heart sounds in the current training set, and to concretely validate the method. Further work include building a new training set recorded from actual patients, then further evaluate the method based on this new training set.

Show MeSH
A heart sound signal and the auto-correlation of its envelope.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3045988&req=5

Figure 9: A heart sound signal and the auto-correlation of its envelope.

Mentions: where N is the signal length. Only positive values of m need to be considered since heart sound signals are real. The length of a cardiac cycle was determined by seeking for the first peak after the origin in the envelope of its autocorrelation function. Since heart sounds are nearly periodic signals, auto-correlation of envelope signals will have peaks where the shifted signal x(n + m) has been shifted by exactly one period and its cardiac cycles are aligned with those of the unshifted x(n). Figure 9 shows a heart sound signal in the top panel and the auto-correlation of its envelope signal in the bottom panel. The waveform in the bottom panel was searched for a maximum within the domain of 1000 to 5000 samples from the beginning of the signal. At fs = 4000 Hz this domain corresponds to heart rates of 48 to 240 beats per minute (BPM), which account for usual human heart rates. The BPM information also allows for identification of abnormal heart rates. The peak marked by a circle is the maximum sample. It can be seen that the distance (number of samples) from this sample to the start of the signal is roughly the same as the length of the first cardiac cycle in the top panel. The assumption is that heart rate remains constant throughout the entire signal.


A framework for automatic heart sound analysis without segmentation.

Yuenyong S, Nishihara A, Kongprawechnon W, Tungpimolrut K - Biomed Eng Online (2011)

A heart sound signal and the auto-correlation of its envelope.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3045988&req=5

Figure 9: A heart sound signal and the auto-correlation of its envelope.
Mentions: where N is the signal length. Only positive values of m need to be considered since heart sound signals are real. The length of a cardiac cycle was determined by seeking for the first peak after the origin in the envelope of its autocorrelation function. Since heart sounds are nearly periodic signals, auto-correlation of envelope signals will have peaks where the shifted signal x(n + m) has been shifted by exactly one period and its cardiac cycles are aligned with those of the unshifted x(n). Figure 9 shows a heart sound signal in the top panel and the auto-correlation of its envelope signal in the bottom panel. The waveform in the bottom panel was searched for a maximum within the domain of 1000 to 5000 samples from the beginning of the signal. At fs = 4000 Hz this domain corresponds to heart rates of 48 to 240 beats per minute (BPM), which account for usual human heart rates. The BPM information also allows for identification of abnormal heart rates. The peak marked by a circle is the maximum sample. It can be seen that the distance (number of samples) from this sample to the start of the signal is roughly the same as the length of the first cardiac cycle in the top panel. The assumption is that heart rate remains constant throughout the entire signal.

Bottom Line: Geometric mean was used as performance index.Average classification performance using ten-fold cross-validation was 0.92 for noise free case, 0.90 under white noise with 10 dB signal-to-noise ratio (SNR), and 0.90 under impulse noise up to 0.3 s duration.Further work include building a new training set recorded from actual patients, then further evaluate the method based on this new training set.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Communication and Integrated Systems, Tokyo Institute of Technology, Japan 2-12-1-W9-108 Ookayama, Meguro-ku, Tokyo, Japan. toey123@gmail.com

ABSTRACT

Background: A new framework for heart sound analysis is proposed. One of the most difficult processes in heart sound analysis is segmentation, due to interference form murmurs.

Method: Equal number of cardiac cycles were extracted from heart sounds with different heart rates using information from envelopes of autocorrelation functions without the need to label individual fundamental heart sounds (FHS). The complete method consists of envelope detection, calculation of cardiac cycle lengths using auto-correlation of envelope signals, features extraction using discrete wavelet transform, principal component analysis, and classification using neural network bagging predictors.

Result: The proposed method was tested on a set of heart sounds obtained from several on-line databases and recorded with an electronic stethoscope. Geometric mean was used as performance index. Average classification performance using ten-fold cross-validation was 0.92 for noise free case, 0.90 under white noise with 10 dB signal-to-noise ratio (SNR), and 0.90 under impulse noise up to 0.3 s duration.

Conclusion: The proposed method showed promising results and high noise robustness to a wide range of heart sounds. However, more tests are needed to address any bias that may have been introduced by different sources of heart sounds in the current training set, and to concretely validate the method. Further work include building a new training set recorded from actual patients, then further evaluate the method based on this new training set.

Show MeSH