基于肤色特征和遗传算法的人眼瞳孔定位算法  

Location algorithm for eye pupil based on skin color features and genetic algorithm

作　　者：郝明刚[1] 董秀成[1] 

机构地区：[1]西华大学四川省信号与信息处理重点实验室,成都610039

出　　处：《计算机应用研究》2012年第7期2754-2757,2762,共5页Application Research of Computers

基　　金：四川省信号与信息处理重点实验室基金资助项目(SGXZD0101-10-1);西华大学研究生创新基金资助项目(YCJJ201169)

摘　　要：提出了一种单目摄像头下定位人眼瞳孔的方法,分为人脸区域检测、人眼区域检测、瞳孔中心定位三个阶段。在人脸区域检测阶段,利用人脸的肤色和唇色在不同色度空间下的特性,结合区域增长的方法分割出人脸区域;在人眼区域检测阶段,利用定位出的人脸区域,根据先验知识缩小搜索区域,再结合遗传算法搜索眼部区域;最后利用圆的几何性质定位瞳孔中心。实验结果证明了本算法在复杂背景和头部偏转情况下的有效性。This paper presented a location method for the human eye pupil with single camera,which was divided into three stages,face region detection,eye region detection and pupil center location.In the face region detection stage,combining the method of regional growth,it separated the face region by the space character of face’s skin color and lip color under different chromaticity.In the eye region detection stage,using the face region positioned,according to prior knowledge,it narrowed the search area,and then searched the eye area with the genetic algorithm.Finally,it positioned pupil center by the geometrical properties of round.Simulation results indicate this algorithm is effective in complex background and circumstances of head deflection.

关 键 词：人脸区域 色度空间 区域增长 眼部区域 遗传算法 定位瞳孔 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]