Accurate Fatigue Detection Based on Multiple Facial Morphological Features

Table 3

Four recent studies on the fatigue detection problem. The column “Year” gives the publication year of each study. And the column “Algorithms” summarizes the techniques used in each study. The column “Accuracy” gives the accuracy achieved by the best model on the complete or largest test dataset in each study. The column “Time (ms)” is the processing time of the algorithm on one image.

Authors	Year	Algorithms	Accuracy	Time (ms)
Ma et al.	2017	Two-stream CNN	91.57%	—
Wathiq and Ambudkar	2017	VJ method, SVM	94.32%	70
Wang et al.	2017	Iris tracking, EBF, PERCLOS	80.75%	—
Gao and Wang	2017	Eye feature extracting, SVM	88%	—