梯度估计在眼睛特征提取中的应用

眼睛特征的提取在表情分析、人脸合成和人脸识别等应用中起着非常重要的作用.目前广泛使用的眼睛特征提取算法是基于模板匹配的算法和基于特征点的算法.然而基于模板匹配的算法通常需要进行多个参数的选择,而且匹配过程非常耗时;特征点的算法则存在准确度较低的问题.针对以上问题提出了一种基于梯度信息提取虹膜中心、虹膜外轮廓和眼睛角点等眼睛特征的算法,算法用改进霍夫变换提取虹膜中心和外轮廓.在获得虹膜中心点后,参照中心位置选择眼睛角点候选区域并对其梯度进行分析.依据角点附近梯度分布与其它区域梯度分布不同的性质定位角点.仿真结果表明提出的算法能有效的提取眼睛特征,在测试数据上得到了比其它算法更高的准确率和更快的速度.

Gradient Estimation for Eye Feature Extraction

Eye features extraction is a vital step for many applications such as facial expression analysis,face synthesis and face recognition.Currently,the widely used method is based on both template deformation or feature point.But template deformation based method needs to determine many parameters and the procedure of deformation is time-consuming,while the accuracy of feature point based method is relatively low.Focusing on the problems described above,an algorithm based on gradient information is presented to ...