计算机工程 ›› 2007, Vol. 33 ›› Issue (23): 208-210.doi: 10.3969/j.issn.1000-3428.2007.23.072

• 人工智能及识别技术 • 上一篇    下一篇

一种由粗到精的人眼定位方法

王 婷,杨国胜,申晓华   

  1. (河南大学先进控制与智能信息研究所,开封 475001)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-12-05 发布日期:2007-12-05

Coarse-to-fine Eye Location Method

WANG Ting, YANG Guo-sheng, SHEN Xiao-hua   

  1. (Institute of Advanced Control and Intelligent Information Processing, Henan University, Kaifeng 475001)
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-12-05 Published:2007-12-05

摘要: 为提高人眼定位算法的实时性和抗噪性,提出了一种基于眉眼区域内“凹陷”地形特征点检测和人眼方差滤波器的人眼定位方 法(GPL)。该方法利用Gabor小波变换和直接提取“凹陷”地形特征点技术,在眉眼区域内搜索眼睛候选点,并且利用构造的人眼方差滤波器精确定位人眼。在有噪声和无噪声的人脸图像上进行了比较仿真试验。结果表明,与纯粹基于地形特征匹配的人眼定位算法相比,GPL在定位实时性、准确性和抗噪性方面都有显著提高。

关键词: Gabor小波, 地形特征, 人眼方差滤波器, 人眼定位

Abstract: In order to improve the real-time and anti-noise performance of the eye location method, this paper presents a coarse-to-fine eye location method (Gabor-pit-location, GPL) based on directly detecting “pit” feature in brow-and-eye region and the eye variance filter. With the direct extraction of “pit” topographic features, the candidate’s eyes are searched in brow-and-eye region extracted from the face image by use of Gabor wavelet transform. The prior knowledge of face structure and the eye variance filter are employed to determine the real eye positions in the face image. Comparative experiments are performed on the images with and without noise respectively, and the results show that GPL is better in the real-time, accuracy and anti-noise performance than the eye location method purely based on the terrain feature matching.

Key words: Gabor wavelet, topographic feature, eye variance filter, eye location

中图分类号: