基于肤色检测和灰度特征的人眼定位方法研究  

ON THE METHOD OF EYES LOCALIZATION BASED ON COMPLEXION DETECTION AND GRAY FEATURE

作　　者：冯浩[1] 王冬欣[1] 王宪保[1] 

机构地区：[1]浙江工业大学智能信息系统研究所,浙江杭州310014

出　　处：《计算机应用与软件》2009年第1期234-235,272,共3页Computer Applications and Software

摘　　要：讨论对于摄像头采集的人脸照片的眼睛定位算法。该算法首先基于肤色检测,在自然背景下检测到人脸的位置,再根据人眼在人脸中具有最大复杂度,结合眼部的低灰度特点,快速地找到人眼位置。方法受人脸表情变化影响较小,快速的特点也完全符合实时系统的需要。实验结果表明,此算法对于复杂背景下采集的人脸照片可以快速准确地检测出人脸并定位人眼,对于小幅度偏转的人脸照片同样适用。人眼正确定位率可达到94%。This paper presents the eyes location algorithm for the face photo collected by the camera. The algorithm is based on complexion detection and firstly detects the position of face from the natural background. Then ,by combining the eyes characters of highest degree in com- plexity on the face and lower grey, it locates eyes position quickly. The face expression changes have little influence on the algorithm, and its fast feature also meets the demand of real time application well. The experimental results show that this algorithm can detect face from the face picture collected in the complex background efficiently and locate the eyes accurately and quickly. It is also suitable for the face photos with small deflection. The correct locating rate of the eyes could reach 94%.

关 键 词：肤色检测 灰度值 人眼定位 复杂度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP391.9[自动化与计算机技术—计算机科学与技术]