A Rotation Insensitive Method Of Positioning In The Face Image In The Eyes

In this paper, a method of locating eyes in a face image is presented. Compared with most methods in this field, this method is insensitive to the rotation of the image. And when applying this method to the color image , eyes can be detected in a profile face image although with some false alarms.This method includes two main steps: drawing the contour of the face and locating the eyes in the face. In this method an improved snake model is used in the first step and a projection based on the polar coordination is employed in the second step. The two methods used in the two steps are all insensitive to the rotation so that this method of locating eyes is also insensitive to the rotation.When dealing with grey-level images, some geometry features related to eyes are generalized to locate eyes more precisely . When dealing with color images , a color transforming is utilized to detect the faces and scleras.In the end of the paper, the method is tested on 80 gray-level images and 44 color images which are selected from the Stirling image database. When applied to grey-level images, this method has a detection rate of 92. 5% and when applied to color images, this method also has a high detection rate but with a lot of false alarms.To test this method' s constancy to the rotation, 160 gray-level images derived from 40 gray-level upright image are tested. In the all 200 images, there are only 5 of them inconstancy. So this method' s constancy to the rotation is 97. 5%.