Drowsiness Detection Using Kapur Segmentation Method

In this paper, we propose a drowsiness detection method using a high-performance Kapur segmentation method, adapted for one-dimensional signals. A FaceMesh algorithm is used to build a 3D model of the face and estimate the degree of eyelid opening for the right eye. The obtained signals are filtered with an EMA (Exponential Moving Average) filter to increase the accuracy of fatigue detection. (Abstract)