Table 2.

Classification results for proposed method using different pre-trained CNNs and dDTF images on MAHNOB-HCI database in all frequency bands while using tenfold Cross-Validation method

Net	Bands	Accuracy (%)	Precision (%)	Recall (%)	F1-score (%)
AlexNet	Delta	79.20 ± 0.75	79.38 ± 0.73	79.41 ± 0.70	79.25 ± 0.71
	Theta	79.45 ± 0.74	79.63 ± 0.72	79.72 ± 0.71	79.53 ± 0.72
	Alpha	79.40 ± 0.73	79.44 ± 0.72	79.38 ± 0.69	79.40 ± 0.70
	Beta	79.56 ± 0.73	79.74 ± 0.73	79.83 ± 0.67	79.40 ± 0.71
	Gamma	79.50 ± 0.74	79.71 ± 0.73	79.80 ± 0.66	79.62 ± 0.68
VGG-19	Delta	86.12 ± 0.75	86.10 ± 0.74	86.35 ± 0.70	86.16 ± 0.67
	Theta	86.45 ± 0.74	86.42 ± 0.73	86.75 ± 0.73	86.39 ± 0.68
	Alpha	88.30 ± 0.75	88.29 ± 0.74	88.30 ± 0.72	88.30 ± 0.72
	Beta	87.54 ± 0.73	87.75 ± 0.72	87.65 ± 0.73	87.56 ± 0.68
	Gamma	87.50 ± 0.74	87.70 ± 0.72	87.61 ± 0.72	87.76 ± 0.65
ResNet-50	Delta	97.30 ± 0.62	97.35 ± 0.66	97.20 ± 0.67	97.31 ± 0.69
	Theta	98.25 ± 0.63	98.50 ± 0.63	98.25 ± 0.65	98.47 ± 0.62
	Alpha	99.43 ± 0.58	99.42 ± 0.61	99.44 ± 0.62	99.42 ± 0.56
	Beta	98.44 ± 0.59	98.67 ± 0.62	98.42 ± 0.63	98.54 ± 0.58
	Gamma	98.35 ± 0.57	98.55 ± 0.65	98.37 ± 0.63	98.29 ± 0.60
Inception-v3	Delta	97.30 ± 0.62	97.26 ± 0.65	97.36 ± 0.65	97.33 ± 0.59
	Theta	97.30 ± 0.63	97.27 ± 0.64	97.36 ± 0.63	97.35 ± 0.61
	Alpha	98.55 ± 0.59	98.55 ± 0.63	98.55 ± 0.64	98.55 ± 0.59
	Beta	97.32 ± 0.60	97.26 ± 0.63	97.39 ± 0.65	97.30 ± 0.60
	Gamma	97.25 ± 0.59	97.29 ± 0.61	97.31 ± 0.60	97.26 ± 0.62

Highest results are showed bold