Abstract
Blinking can be used as a biological indicator of consciousness, concentration power and Visual Load. We develop a technique to obtain the number of blinking during the VDT work easily and without restrictive measurement method. In the technique, images from a generally utilized Web camera connected to PC are processed to detect the blinking. The template matching method for the image is used to track the eye region and the upper eyelid which moves up and down when blinking. In the previous paper, the accuracy of blinking detection was over 90% at the capture resolution 320×240 [pixel]. However, when the capture resolution was set even lower, the tracking accuracy of the upper eyelid inside the eye region deteriorated extremely. In order to improve the tracking accuracy, the continuous change of the number of pixels which corresponds to the size of pupil and iris within the eye region is used to detect the movement of the upper eyelid. Therefore we examined whether the fatigue estimation of VDT work by measured Eye-Blink.
