Skip to main content

Table 3 Time-frequency domain

From: Localization and classification of heart beats in phonocardiography signals —a comprehensive review

Study

Year

Pre-processing techniques

Localization techniques

Classification techniques

Beat types

Data set

Signals statistics

Evaluation metrics

Results

Zhang et al. [52]

2017

Resampling, filtering, and normalization

Time, spectrogram, and regression

SVM

Normal and abnormal heart sounds

PASCAL CHSC 2011

176 dataset A and 507 dataset B samples

Precision, sensitivity, and specificity

60–76%, 89–100%, and 39–90%

Zhang et al. [53]

2017

Resampling, filtering, and normalization

Time, tensor representation and, feature extraction

SVM

Normal and abnormal heart sounds

PASCAL CHSC 2011 and 2016 PhysioNet challenge

176 dataset A and 507 dataset B samples

Precision, sensitivity, and specificity

25–67%, 44–100%, and 39–64%

Singh et al. [50]

2017

Time, probability, and spectrogram- based processing

Time, probability, and spectrogram-based processing

Normal and abnormal heart sounds

National Institute of Health, India

Signal recoding from 45 healthy and 45 patients

Barma et al. [51]

2016

S3 separation and decomposition by HVD

TFR by SPWVD

Time duration-based classification

82 cycles

Data collected from Texas Heart Institute and Michigan Heart Sound and Murmur Library dataset

Efficiency

93.9%

Gavrovska et al. [48]

2014

Signal decimation

Pseudo-affine Wigner-Ville distribution and Haar wavelet lifting, autocorrelation, and Shannon envelogram

Pseudo-affine Wigner-Ville distribution and Haar wavelet lifting, autocorrelation, and Shannon envelogram

Normal heart sounds

Data acquisition at University Children Hospital in Belgrade, Serbia

Signal recoding from 35 healthy and 55 patients

Recall and precision

90.41 and 96.39%

Boutana et al. [46]

2011

Spectrogram with Reyni marginal entropy

Spectrogram with Reyni marginal entropy

Normal heart sounds, early aortic stenosis, late aortic stenosis, pulmonary stenosis, and mitral regurgitation

E-General Medical

6 signals

Debbal et al. [47]

2005

STFT, CWT, and WD

STFT, CWT, and WD

Normal hear t sounds, mitral stenosis, split S2, aortic insufficiency, and stenosis

-

-

-

-

Zhang et al. [45]

1998

MP-based Wigner distribution

MP-based Wigner distribution

Normal heart sounds and murmurs

Private dataset

11 signals

Error reduction