Automatic Detection of Obstructive Sleep Apnea Events Using a Deep CNN-LSTM Model

Table 6

Comparison of performances of different methods.

Input	Author	Method	TAC (%)	SE (%)	SP (%)
ECG	Jafari [43]	Handcrafted features, SVM	94.8	94.1	95.4
	Chen et al. [44]	Handcrafted features, SVM	82.1	83.2	80.2
	Urtnasan et al. [25]	CNN	96	96	96
	Banluesombatkul et al. [34]	CNN	79.45	77.6	80.1
	Zarei and Asl [12]	Handcrafted features, SVM	94.63	94.43	94.77
	Tripathy [45]	Handcrafted features, kernel extreme learning machine	76.37	78.02	74.64
	Hassan and Haque [46]	Handcrafted features, RUSboot	88.88	87.58	91.49
	Hassan [47]	Handcrafted features, AdaBoost	87.33	81.99	90.72
	Our method	CNN	96.1	96.1	96.2