基于Hough变换的人眼检测方法研究  

Research on detection of human eye based on Hough transfor m

作　　者：吴冰[1] 陈步威[1] 李富贵[1] 杨文强[1] 

机构地区：[1]河南理工大学电气工程与自动化学院,河南焦作454003

出　　处：《河南理工大学学报（自然科学版）》2008年第5期556-559,564,共5页Journal of Henan Polytechnic University（Natural Science）

基　　金：河南省教育厅自然科学研究计划项目(2008B510009)

摘　　要：人脸检测在许多应用中都是重要的一个处理阶段,眼睛是一个在人脸检测中极为重要的人脸特征,因此一种精确可靠的定位眼睛的算法对许多实际的应用是十分重要的.提出一种基于Hough变换椭圆检测的人眼睛定位方法,先对人脸图象进行canny边缘提取,然后根据眼睛的生理特征,利用提取的眼睛边缘特征进行椭圆检测.实验证明该方法定位准确,能有效的抑制噪声,并且有很高的识别率.Face detection is an important step in many applications such as human face recognition, video conferencing, human computer interfacing etc. The eye is the most important facial feature in face detection. Thus, a reliable and fast method for locating the eyes precisely in an image is vital to many practical applica- tions. A new method for detection and recognition of human eye is proposed in the paper, based on ellipse and Hough transform. At first, images of the human face edge are extracted based on canny method. Then, according to the physiological characteristics of the eye, ellipse detection is done from the eyes of extracting edge features. The results show that this method has positioning accuracy, effective noise suppression, and a high recognition rate.

关 键 词：HOUGH变换 椭圆检测 眼睛检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]