Analysis of eye closure duration based on the height of iris

Eye closure duration is one of the parameters that is regularly chosen in detecting the state of drowsiness of a person. In order to analyze the eye closure duration, a process involving eye state classification is done to classify whether the eye is in an open state or in a closed state. This paper introduces a new parameter which is the height of iris to classify the eye state and to analyze the eye closure duration. Face and eye detection is the first step and Viola-Jones algorithm is implemented in the procedure. Next, face and eye tracking is done by utilizing the Kanade Lucas Tomasi (KLT) algorithm which tracked the feature points. The extracted eye region is further used to localize the iris through image enhancement process. The most crucial step in eye closure duration is the eye state classification process. The eye state is classified based on the height of the iris that has been localized. The height of the iris is acquired from the bounding box drawn surrounding the area of the localized iris. Finally, after the state of an eye is classified successfully, the eye closure duration is analyzed through the plotted graph between the heights of the iris against a period of time. Furthermore, by evaluating the iris' height, the size of an eye for each subject is classified. The proposed algorithm is implemented on Zhejiang University (ZJU) database.