Abstract: Whether a driver is fatigued can be reflected by his eyes during the course of driving，so it is practical to use the information about the driver’s eyes to monitor driver’s fatigue state and among this，eye location is the key problem.This paper proposes a new method for locating the eyes precisely in face detection.The new method consists of two main parts.In first part，we use skin cluster algorithm to divide human face region.In skin segmentation graph we do geometry filter.We calculate the centroid of hole in the face region to find human eyes.In second part，this paper proposes a new eye model.A new ellipse detection approach based on Hough transform is applied to locate the irises precisely according to the face region and eye region.Experimental results show that the proposed approach is fast and reliable.

摘要： 通过人眼图像来检测驾驶员疲劳驾驶是目前的主流方向，面部及眼睛定位是其中关键的环节。提出了一种新颖的精确定位眼睛的方法。该方法由两部分组成：第一部分，通过肤色聚类分割算法将人脸区域分割，对分割图进行几何过滤，对得到的候选人脸区域中的孔洞计算质心点找到可能的人眼对；第二部分，在检测到人脸区域和眼睛大致位置的基础上，结合提出的眼睛模型，采用新的Hough变换椭圆检测算法精确定位人眼的位置。实验证明所提出的算法是快速可靠的。