| 基于多特征融合的人脸表情识别 |
| |
| 引用本文： | 朱娅妮,杜加友. 基于多特征融合的人脸表情识别[J]. 杭州电子科技大学学报, 2009, 29(5) |
| |
| 作者姓名： | 朱娅妮 杜加友 |
| |
| 作者单位： | 杭州电子科技大学科技处,浙江,杭州,310018 |
| |
| 基金项目： | 杭州电子科技大学科学研究基金资助项目 |
| |
| 摘 要： | 该文提出一种结合均值主元分析和线性判别进行特征选择的人脸表情识别方法。首先对训练样本做均值主元分析,然后再做线性判别,构造特征子空间,最后用支持向量机的方法进行分类。在美国耶鲁大学人脸表情库上分别进行了测试人不参与训练仿真实验,并与其他方法进行比较,该文提出的方法取得了更好的识别结果。
|
| 关 键 词： | 表情识别 均值主元分析 线性判别 支持向量机 |
Facial Expression Recognition Based on the Multi-Feature Fusion |
| |
| ZHU Ya-ni; DU Jia-you. Facial Expression Recognition Based on the Multi-Feature Fusion[J]. Journal of Hangzhou Dianzi University, 2009, 29(5) |
| |
| Authors: | ZHU Ya-ni DU Jia-you |
| |
| Affiliation: | ZHU Ya-ni; DU Jia-you(D epo.Sf ciencea ndT echnologyR esearchH; angzhouD ianzUi niversityH; angzhouZ hejiang310018; China); |
| |
| Abstract: | In this paper,a novel method of expression recognition is presented.Our method uses the equable prin-cipal component analysis(EPCA) and fisher’s linear discriminate(FLD) as face representation and employs support vector machine(SVM) as face classifiers.Experiments of human who has not been trained are performed on the Yale face database.Compared to other methods,our method can get higher recognition ratio.Therefore,it is feasible to apply EPCA-FLD-SVM to expression recognition. |
| |
| Keywords: | expression recognition equable principal component analysis fisher line discriminate support vector machine |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
