Driver Fatigue Detection Method Based on Human Pose Information Entropy

Table 2

Fatigue detection results among different classifiers.

Classifier	Setting	Accuracy (%)		Precision (%)		Recall (%)
		Train set	Test set	Train set	Test set	Train set	Test set
BAYES	—	90.13	89.36	88.51	87.56	91.63	88.53
MLP	(32, 32)	96.3	98.1	95.51	98.23	98.62	97.53
	(64, 64)	96.4	97.85	95.68	98.12	98.31	97.43
	(128, 128)	96.7	98.1	94.64	98.21	98.73	97.45
	(32, 32, 32)	98.32	98.01	97.91	97.42	98.83	99.21
	(64, 64, 64)	98.46	98.13	97.93	97.47	98.87	99.13
	(128, 128, 128)	98.70	98.30	98.11	97.50	98.98	99.23
SVM	T = 1, d = 10	95.07	95.75	93.39	94.32	96.95	97.20
	T = 1, d = 8	94.30	94.22	92.87	93.71	95.93	94.91
	T = 0	98.81	99.35	99.23	98.98	98.61	99.49