J4 ›› 2009, Vol. 47 ›› Issue (05): 999-1005.

• 计算机科学 • 上一篇    下一篇

一种基于准二值图的眼仁定位新方法

 高岩, 赵丽娜   

  1. 北京化工大学 理学院, 北京 100029
  • 收稿日期:2008-10-07 出版日期:2009-09-26 发布日期:2009-11-03
  • 通讯作者: 赵丽娜 E-mail:zhaoln@mail.buct.edu.cn.

A Human Eye Location Algorithm Based onQuasi Binary Image

GAO Yan, ZHAO Lina   

  1. College of Science, Beijing University of Chemical Technology, Beijing 100029, China
  • Received:2008-10-07 Online:2009-09-26 Published:2009-11-03
  • Contact: ZHAO Lina E-mail:zhaoln@mail.buct.edu.cn.

摘要:

提出一种基于以二次二值化图像为研究对象, 二次垂直投影为定位手段的眼睛定位方法. 先用小波8层分解原图像, 垂直和水平投影定出眼部区域的坐标; 再对原图像进行奇异值分解, 取最大奇异值的1/3.5为新值, 对图形进行重建得到对眼仁有较好刻画的准二值图. 用已知的眼部坐标在无噪声准黑白图上截取眼睛部分后, 再次用垂直投影方法得到眼仁的准确位置.

关键词: 奇异值分解; 小波分解; 投影函数; 人眼定位

Abstract:

A human eye localization method is described in this paper. It’s based on twice binary images and twice vertical projection. Firstly, the original image is decomposed with 8level of wavelet, and then the main area for eyes is gotten by vertical and horizontal projection. Secondly, the given image is decomposed with SVD again and  a  max singular value of 1/3.5 is set to be a new threshold and to reconstruct the original image for obtaining the quasi binary image which embodies eyes better. At last, eyes area is intercepted from lessnoise quasi binary image by already known coordinates and finally, the accurate position of eyes is achieved by vertical projection once again.

Key words: SVD, wavelet decompose, projection function, eye location

中图分类号: 

  • TP391