Table 8.

Benchmarking with State-of-the-Art CNNs.

Method	Class	Data Validation	Performance (%)
			Accuracy	Sensitivity	Specificity
Ensemble Neural Network [9]	2 classes (normal vs. and HLHS)	intra-patient	-	89	92
	2 classes (normal vs. TOF)	intra-patient	-	71	89
Residual learning [13]	2 classes (normal vs. CHDs including HRHS, HLHS, highly RAS)	intra-patient	93	93	-
	2 classes (normal vs. CHDs including HRHS, HLHS, highly RAS)	inter-patient	91	91	-
Deep learning model [14]	2 classes (normal vs. TOF)	intra-patient	-	75	76
	2 classes (normal vs. HLHS)	intra-patient	-	100	90
DGACNN [15]	2 classes (normal vs. CHD)	intra-patient	85	-	-
Proposed	2 classes (normal vs. CHDs including ASD, VSD, AVSD, EA, TOF, TGA, HLHS)	intra-patient	100	100	100
	2 classes (normal vs. CHDs ASD, VSD, AVSD, EA, TOF, TGA, HLHS)	inter-patient	92	91	92
	8 classes (normal, CHDs ASD, VSD, AVSD, EA, TOF, TGA, HLHS)	inter-patient before augmentation	71	62	68
	8 classes (normal, CHDs ASD, VSD, AVSD, EA, TOF, TGA, HLHS)	Inter-patient after augmentation	99	97	98