Eye positioning based on windowed gray-scale integral projection algorithm

Abstract

Eye positioning is an important prerequisite for vehicle-mounted fatigue driving monitoring. In this paper, a windowed gray-scale integral projection algorithm based eye-positioning method is introduced. Firstly, the human eye is positioned by Adaboost algorithm, then the eyebrows are separated by the windowed gray-scale integral projection algorithm, consequently, the accurate positioning of human eye parts is achieved. The experimental results show that the separation effect is obvious and the speed of human eye positioning is significantly improved.