Deep Layer Kernel Sparse Representation Network for the Detection of Heart Valve Ailments from the Time-Frequency Representation of PCG Recordings

Table 6

Summary of automated detection of HVA developed using PCG signals using the same database.

Methods used for feature extraction	Classifiers	Classes	Accuracy (%)
Morphological features extracted from PCG recording [77]	SVM	N, MS, MR	91.23
Wavelet entropies as features from PCG [86]	ANFIS	N, PS, MS	98.33
Multilevel basis selection- (MLBS-) based wavelet features extracted from PCG [76]	SVM	N, AS, MR, AR	97.56
Entropy and energy fraction-based features [78]	SVM	N, TI, PS, MI, MS	97.17
Wavelet and MFCC features obtained from PCG [44]	SVM	N, AS, MS, MR, MVP	97.90
Magnitude and phase features extracted using SST of PCG [36]	Random forest	N, AS, MS, MR	95.13
Features extracted using CT of PCG [13]	Multiclass composite classifier	HC, AS, MS, MR	98.33
DNN [79]	WaveNet	N, MS, MR, AS, MVP	98.20
Proposed work (features evaluated in SCT domain of PCG)	DLKSRN	N, MS, MR, AS, MVP	99.24