Table 5.

Evaluation metrics on the test set.

Model	Class Name	Precision	Recall	F1-Score	Accuracy
Inception V3	Not drowsy	0.9895 ± 0.006	0.9896 ± 0.006	0.9895 ± 0.002	0.9895 ± 0.002
	Drowsy	0.9897 ± 0.005	0.9894 ± 0.006	0.9895 ± 0.002
VGG16	Not drowsy	0.9868 ± 0.001	0.9798 ± 0.006	0.9833 ± 0.003	0.9833 ± 0.003
	Drowsy	0.9800 ± 0.006	0.9869 ± 0.001	0.9834 ± 0.003
ResNet50V2	Not drowsy	0.9967 ± 0.003	0.9931 ± 0.003	0.9949 ± 0.003	0.9949 ± 0.003
	Drowsy	0.9932 ± 0.003	0.9967 ± 0.003	0.9949 ± 0.003
Proposed DD-AI	Not drowsy	0.9988 ± 0.001	0.9988 ± 0.001	0.9988 ± 0.001	0.9988 ± 0.001
	Drowsy	0.9988 ± 0.001	0.9988 ± 0.001	0.9988 ± 0.001