基于彩色分割的虹膜检测

Iris Detection Based on Color Segmentation

提出了一种基于彩色分割的虹膜检测方法,对于一幅经过准确定位的眼睛窗口图像,它的彩色信息比灰度信息更有助于进行虹膜检测。首先利用图像颜色的饱和度信息将眼睛图像中的眼睛区域与皮肤区域分离,然后利用亮度信息将眼睛区域的眼白和虹膜分离得到虹膜区域,再通过Hough变换进行虹膜检测。结果证明该方法不仅检测率高,而且能适应一定人脸姿势、眼睛视线方向的变化,甚至在眼睛区域较暗的情况下,也能很好地检测到虹膜。

This paper proposes an iris detection method based on color segmentation. To a localized eye image, color images contain more information helpful for detecting irises than intensity images. First, the saturation of color is used to divide the eye and skin regions from the eye image. Then, the eye region is divided into the iris and the white of the eye using the brightness of color. Finally, the position of the iris is determined by the Hough transform. The outcome shows that the method not only achieves an high accuracy rate, but it can adaptable to individual variation and directions of faces and eves. The iris can be detected even if the eye regions are dark.