Interpatient ECG Heartbeat Classification with an Adversarial Convolutional Neural Network

Table 3

Record-level performance of the proposed method on DS2.

	Number of heartbeats				N		S		V		Overall
Recording	All	N	S	V	Sen	Pre	Sen	Pre	Sen	Pre	Acc
100	2264	2231	32	1	100.0%	99.3%	53.1%	100.0%	100.0%	100.0%	99.3
103	2075	2073	2	0	99.7%	99.9%	0.0%	0.0%	—	—	99.6
105	2563	2517	0	41	88.6%	99.1%	—	—	56.1%	42.6%	87.9
111	2115	2114	0	1	100.0%	100.0%	—	—	100.0%	100.0%	100.0
113	1786	1780	6	0	100.0%	99.7%	0.0%	0.0%	—	—	99.7
117	1526	1525	1	0	100.0%	100.0%	100.0%	100.0%	—	—	100.0
121	1854	1852	1	1	99.8%	99.9%	0.0%	0.0%	100.0%	25.0%	99.8
123	1509	1506	0	3	95.5%	100.0%	—	—	100.0%	100.0%	95.5
200	2592	1739	30	821	99.4%	95.1%	0.0%	0.0%	92.0%	98.7and	95.8
202	2127	2052	55	19	86.9%	98.3%	27.3%	100.0%	94.7%	66.7%	85.4
210	2561	2415	22	194	96.0%	97.4%	27.3%	28.6%	74.2%	100.0%	93.4
212	2739	2739	0	0	100.0%	100.0%	—	—	—	—	100.0
213	3242	2632	28	220	99.8%	89.2%	0.0%	0.0%	77.3%	94.4%	88.7
214	2253	1995	0	255	61.0%	99.7%	—	—	99.2%	98.8%	65.2
219	2145	2073	7	64	99.1%	99.4%	0.0%	0.0%	92.2%	77.6%	98.6
221	2418	2024	0	394	100.0%	99.8%	—	—	98.7%	100.0%	99.8
222	2474	2265	209	0	98.2%	94.7%	40.7%	68.5%	—	—	93.3
228	2044	1679	3	362	95.1%	98.3%	0.0%	0.0%	92.5%	81.5%	94.5
231	1563	1560	1	2	100.0%	99.9%	0.0%	0.0%	50.0%	100.0%	99.9
232	1772	395	1377	0	99.0%	87.7%	95.6%	99.8%	—	—	96.4
233	3070	2225	7	827	99.7%	98.9%	28.6%	100.0%	98.1%	98.7%	98.8
234	2744	2691	50	3	100.0%	98.2%	0.0%	0.0%	100.0%	100.0%	98.2
Total	49516	44082	1831	3208	96.2%	98.0%	78.8%	90.8%	92.5%	94.3%	94.7