Extraction of Individual Feature in Eye Blinks by Using Normalized Correlation

Abstract

We developed an individual identification method using the difference of blinking speed distribution. This feature of blinking was measured by a high-speed camera and extracted the time variation as a percentage of the eyelid aperture. Next, a calculation of normalized correlation was used to evaluate the speed distribution in each position of eyelid. As a result of comparative experiment between voluntary and spontaneous brink by 6 participants, we found that the accuracy of individual identification with voluntarily is better than spontaneity.