Driver Fatigue Detection Method Based on Human Pose Information Entropy

Table 6

Comparison of the proposed method with other methods.

Methods	Feature	Feature types	Approach	Accuracy (%)
Du et al. [33]	Information fusion	Heart rate, eye openness level, and mouth openness level	Multimodal fusion recurrent neural network	91.67
Liu et al. [45]	Facial features	Closed eye and yawning	Fuzzy inference system and PERCLOS	96.5
Zhao et al. [46]	Expressions	Driver fatigue expression	Deep belief network (DBN)	96.7
Ansa et al. [34]	Head posture	Yawning, nodding, and shaking	New modified reLU-BiLSTM deep neural network	97.6
Ours	Human pose	Euclidean distance between arms, the area between arms, dispersion of wrist coordinate	Information entropy and SVM	99.35