Intelligent Diagnosis of Heart Murmurs in Children with Congenital Heart Disease

Table 1

Performance of the current and previous methods in pediatric PCG segmentation or classification.

Method	Segmentation or classification type	Performance
Wavelet transform, singular value decomposition [11]	S1 and S2	Acc = 92.1%
Hilbert envelope, ANN [15]	Normal and six pathologies	Se = 91%, Sp = 94%
Matched filters, support vector machine, ANN [23]	Innocent Still’s murmur and other murmurs	Se = 84%∼93%; Sp = 91%∼99%
Shannon energy, wavelet transform [24]	S1 and S2	Acc = 88%
Current method	S1, S2, CAV, CSV, normal, and CHD	Acc = 93%, Se = 93.5%, Sp = 91.7%