Atrial Fibrillation Beat Identification Using the Combination of Modified Frequency Slice Wavelet Transform and Convolutional Neural Networks

Table 5

Comparison with reference studies.

Algorithm	Data	Acc (%)	Se (%)	Sp (%)	AUC
Chen et al. [11]	50 signals	—	89.68	86.16	—
Maji et al. [12]	178 cycles	—	90.48	92.89	—
Ladavich and Ghoraani [13]	14,600 beats	—	90.67	88.87	—
Proposed method	All recordings	81.07	74.96	86.41	0.88
Proposed method	All recordings but excluding one fold with extreme noisy recording	84.85	79.05	89.99	0.92