Interpretable Detection and Location of Myocardial Infarction Based on Ventricular Fusion Rule Features

Table 12

Comparison of the proposed method with other related literature.

Reference	Class number	Feature	Classifier	Performance (%)
Lin et al. [5]	2	MODWPT, statistical	KNN	Acc = 99.57; Se = 99.82; Sp = 98.79
Baloglu et al. [12]	11	End-to-end	CNN	Acc = 99.78;
Han and Shi [13]	7	End-to-end	ResNet	Acc = 99.72; Se = 99.63; Sp = 99.72;
Han and Shi [28]	2	MODWPT, morphological	SVM	Acc = 99.81; Se = 99.56; +p = 99.74
Acharya et al. [29]	11	DWT	KNN	Acc = 98.80; Se = 99.45; Sp = 96.27
Padhy and Dandapat [47]	6	Singular Value Decomposition	SVM	Acc = 95.30; Se = 94.60; Sp = 96.00
Liu et al. [48]	6	End-to-end	CNN	Acc = 99.81
Proposed	10	DWT, rule features	XGBoost	Acc = 99.86; Se = 99.86; Sp = 99.86