ECG Heartbeat Classification Based on an Improved ResNet-18 Model

Table 7

Sensitivity (Se) and precision (P+) of the improved ResNet-18 model, in comparison with state-of-the-art models, for AAMI heartbeat classes V and S.

Model (ensemble learning is omitted from this comparison as there are neither sensitivity nor precision data presented for it in [11])	AAMI heartbeat class V		AAMI heartbeat class S
	Se (%)	P+ (%)	Se (%)	P+ (%)
BbNNs [18]	86.60	93.30	50.60	67.90
End-to-end DNN [15]	93.70	Not provided	77.30	Not provided
1D-CNN [10]	93.90	90.60	60.30	63.50
Improved ResNet-18 (the proposed model)	93.83	97.44	76.45	83.87