Anomaly detection by using a combination of generative adversarial networks and convolutional autoencoders

EURASIP Journal on Advances in Signal Processing

Table 3 Comparative experiments

Metric	Method	ECG1	ECG2	2D gesture
AUROC	BeatGAN	0.7056	0.7329	0.7256
	RAE	0.7502	0.8289	0.7601
	RRN	0.7623	0.7405	0.7530
	RAE-ensemble	0.7788	0.8570	0.7808
	USAD	0.7223	0.4845	0.4852
	Our method	0.8058	0.8904	0.8048
	BeatGAN	0.4101	0.2254	0.4952
	RAE	0.4249	0.4996	0.4979
AUPRC	RRN	0.5653	0.4139	0.4866
	RAE-ensemble	0.4769	0.5256	0.5287
	USAD	0.3470	0.2289	0.2281
	Our method	0.8669	0.5942	0.7697
	BeatGAN	0.4204	0.2931	0.4941
	RAE	0.4736	0.5046	0.5300
F1	RRN	0.5502	0.4537	0.5240
	RAE-ensemble	0.5016	0.5333	0.5511
	USAD	0.3633	0.4124	0.4119
	Our method	0.7635	0.6539	0.7314