Eye state detection based on nonnegative sparse coding feature

It is very useful to detect our eye open or close state in some sistuations. For example, it will give much help to the drivers for driving with drowsiness. In this paper, we propose an efficient method which is made up of four steps for eye state detection: Eye Localization, Preprocessing, Nonnegative sparse coding feature, SVM classifier. We compared our method to other methods[6][7] in our collected dataset which contains the eyes with glasses or in darkness, and so on. In our method, we will preprocess the image for eliminating the illumination and then analyze the eye image feature based on nonnegative sparse coding. It is proved to be very effective in our dataset.