Eye Detection Based on Integral Projection and Hough Round Transform

摘要

As the judgment of driver's fatigue, we must first achieve the driver's face detection, and then conduct the human eye's detect. This work is in color model and radial basis function networks realize the driver's face detection, next using integral projection and Hough round Transform comply eye detection. This method have two advantages, the first is use of integral projection function can achieve coarse positioning of the human eye and be united in wedlock of Hough round Transform finish the exact position of human eyes, so we can make up the shortcomings inaccurate of one single integral projection template matching methods to detect, another advantage is in order to improve the detect accurate we take the human eye as a whole, so we may quickly achieve the human eye detect, thus the errors of monocular detection will have greatly improve.