An Automatic System for Atrial Fibrillation by Using a CNN-LSTM Model

Table 2

Comparison with previous work on the MIT-BIH AF database.

Author	Database	Features	Classifier	Accuracy (%)	Sensitivity (%)	Specificity (%)
Xu et.al. [15]	AFDB	MESWT	CNN	85.85	79.05	89.99
Wei et.al. [14]	AFDB	RCN	CNN	94.59	94.28	94.91
Andersen et.al. [11]	AFDB	RRI	LSTM + CNN	87.4	98.6	86.4
	MITDB
	NSRDB
Dang et.al. [17]	AFDB	RR	CNN-BLSTM	96.59	99.93	97.03
		P-QRS-T	CNN-LSTM	94.07	94.25	92.73
Proposed model	AFDB	Deep features	CNN-LSTM	97.21	97.34	97.08