一种应用于人脸正面图像的眼睛自动定位算法

李粉兰; 徐可欣

doi:null

您当前的位置：

首页 >

文章列表页 >

一种应用于人脸正面图像的眼睛自动定位算法

更新时间：2020-08-12

- 一种应用于人脸正面图像的眼睛自动定位算法
- An algorithm applied in frontal-view face images for automatically localizing eyes
- 光学精密工程 2006年14卷第2期页码：320-326
- 作者机构：
  
  天津大学精密测试技术及仪器国家重点实验室天津,中国,300072
- 作者简介：
- 基金信息：
- DOI：
  中图分类号： TP391.4
- 收稿日期：2005-11-22，
  
  修回日期：2006-01-18，
  
  网络出版日期：2006-04-30，
  
  纸质出版日期：2006-04-30
- 稿件说明：
移动端阅览
李粉兰, 徐可欣. 一种应用于人脸正面图像的眼睛自动定位算法[J]. 光学精密工程, 2006,14(2):320-326.

LI Fen-lan, XU Ke-xin. An algorithm applied in frontal-view face images for automatically localizing eyes[J]. Optics and precision engineering, 2006, 14(2): 320-326.
李粉兰, 徐可欣. 一种应用于人脸正面图像的眼睛自动定位算法[J]. 光学精密工程, 2006,14(2):320-326. DOI：

LI Fen-lan, XU Ke-xin. An algorithm applied in frontal-view face images for automatically localizing eyes[J]. Optics and precision engineering, 2006, 14(2): 320-326. DOI：

摘要

基于人眼轮廓所具有的椭圆特性

提出了一种眼睛定位算法。首先对人脸图像边缘提取

并对边缘所有点进行最小平方椭圆拟合

根据眼睛宽度与高度比例

给定一个阈值

确定眼睛候选区域;然后运用训练过的支持向量机(SVM)分类器验证眼睛区域;最后在眼睛区域

采用一种基于方向性圆形霍夫变换(Circle Hough Transform (CHT))的检测算子定位眼球中心

并对ORL及自建人脸库进行了测试。结果表明:分别在1.25 s

3.0 s内

验证率为91.6%

95%的情况下

该方法可以获得7.26和3.45个像素的平均定位误差

同时也证实了该方法对人脸的平面内旋转、尺度、表情等变化具有较强的鲁棒性。

Abstract

Based on the ellipse-like outline of human eyes

an algorithm for localizing eyes was presented. After edge extracting from the face image

least square ellipse fitting was performed on all points on the edges. According to the proportions of eyes width and height

a threshold was given to confirm eyes candidate regions. A trained Support Vector Machine (SVM) classifier was applied to identify eyes regions. Finally

a circle detection operator based on directional Circle Hough Transform (CHT) was used to localize the eyeball center in the regions resulting from SVM classifier. The experiments were carried out on ORL and self-built face database. The results indicate that the method can accomplish the localization in 1.25 s and 3.0 s respectively

while accurate identification can reach 91.6% and 95%

the average localization errors are 7.26 and 3.45 pixels respectively. At the same time

the results also demonstrate that the method is very robust to face's in-plane rotation

scale and facial expression variation.

关键词

Keywords

references

. 梁毅雄,龚卫国,潘英俊,等. 基于奇异值分解的人脸识别方法[J]. 光学精密工程, 2004,12(5):543-549. LIANG Y X,GONG W G,PAN Y J, et,al. Singular value decomposition-based approach for face recognition[J]. Optics and Precision Engineering, 2004,12(5):543-549.

. 闫丽,袁家虎,刘江,等. 光学子波用于图像处理 . 光学精密工程, 2000, 8(3):225-230. YAN L,YUAN J H,LIU J, et.al. Optical-wavelet transform application in image processing[J]. Optics and Precision Engineering, 2000, 8(3):225-230.

. ROSIN P L. A note on the least squares fitting of ellipses[J]. Pattern Recognition Letters, 1993, 14:799-808.

. GANDER W,GOLUB G H,STREBEL R. Least-square fitting of circles and ellipses[J]. BIT, 1994, 34: 558-578.

. TAUBIN G. Estimation of planar curves, surfaces and non-planar space curves defined by implicit equations, with applications to edge and range image segmentation[J]. IEEE Trans. Pattern Analysis and Machine Intelligence, 1991, 13(11):1 115-1 138.

. BOOKSTEIN F L. Fitting conic sections to scattered data[M]. Computer Graphics and Image Processing, 1979, 9(1): 56-71.

. TAKEUCHI K,COLLIER N. Bio-medical entity extraction using support vector machines . Proc ACL 2003 Workshop on NLP in Biomedicine, Sapporo, 2003.

. JOACHIMS T. Learning to classify text using support vector machines: . Norwell, MA, USA: Kluwer Academic Publishers, 2002.

. ORAZIO T D,LEO M,CICIRELLI G, etal. An algorithm for real time eye detection in face images . Proceedings of the 17th International Conference on Pattern Recognition, Cambridge, 2004.

浏览量

551

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于SURF特征描述符和杰卡德距离的文物碎片拼接

采用核相关滤波器的长期目标跟踪

光学系统弥散斑参数的测试

基于3×3耦合器的双马赫-曾德尔干涉仪数字化相位解调

光电经纬仪CCD曝光中心测量系统的设计