| 一种基于Fisher准则的二维主元分析表情识别方法 |
| |
| 引用本文： | 程剑,应自炉.一种基于Fisher准则的二维主元分析表情识别方法[J].五邑大学学报(自然科学版),2006,20(2):42-46. |
| |
| 作者姓名： | 程剑 应自炉 |
| |
| 作者单位： | 1. 五邑大学,信息学院,广东,江门,529020
2. 五邑大学,信息学院,广东,江门,529020;北京航空航天大学,电子信息工程学院,北京,100083 |
| |
| 基金项目： | 广东省博士启动基金;国家重点实验室基金 |
| |
| 摘 要： | 提出了一种基于Fisher准则进行特征选择的二维主元分析表情识别方法.首先对训练样本做二维主元分析,然后再根据Fisher准则,按Fisher比的大小选择特征向量作为投影轴,最后用最近邻方法进行分类.在JAFFE人脸表情静态图像库上进行实验,与按特征值的大小来选择特征向量相比,该方法更有效.
|
| 关 键 词： | Fisher准则 二维主元分析 表情识别 |
| 文章编号： | 1006-7302(2006)02-0042-05 |
| 修稿时间： | 2005年11月21 |
A Novel Two-Dimensional PCA Facial Expression Recognition Method Based on Fisher Ratio |
| |
| CHENG Jian,YING Zi-lu.A Novel Two-Dimensional PCA Facial Expression Recognition Method Based on Fisher Ratio[J].Journal of Wuyi University(Natural Science Edition),2006,20(2):42-46. |
| |
| Authors: | CHENG Jian YING Zi-lu |
| |
| Abstract: | In this paper,a novel two-dimensional principal component analysis(2-D PCA) method based on Fisher ratio is proposed for facial expression recognition.First,2-D PCA is applied to training sample expression images.Then,the feature vectors are chosen according to their Fisher Ratio and are used as projection vectors.Finally,NN is used as the classifier to classify facial expressions.The results of experiments on JAFFE facial expression database show that the method can improve the recognition rate. |
| |
| Keywords: | Fisher Ratio 2DPCA facial expression recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
