Eye blink based warning system for eye health while using computers

In this paper, we present a computer vision based real-time warning system to prevent the formation of eye diseases of people working continuously in front of computers. This system takes an image of user sitting in front of the computer by web camera then processes it to determine how often user blinks. The system warns user, if there is not sufficient number of eye blinks for a long time. The general flow of the eye blink based warning system consists of three main steps: face and eye detection, feature extraction of preprocessed eye images, and open-close classification of eye images. The system has attained high performance in the experiments. Moreover, the developed system has been turned into a computer plug-in that can work real-time and is available for practical use.