摘要： 先通过PCA特征脸或插值降维,并利用Fisher鉴别矢量集,获得人脸鉴别特征;然后采用常见距离分类器及其组合形式进行识别分析,分类器组合使用多数投票规则和最大值规则等;最后对计算结果进行分析。该文研究思路和方法简洁,结果令人满意,对基于生物特征鉴别分析的工程应用具有较大价值。

关键词: 模式识别, 分类器组合, 人脸特征, 特征提取

Abstract: First, face discriminant features are obtained from the computation of PCA or interpolation methods as well as Fishers methods. Then, face recognition analysis is carried out by six distance classifiers and their combination based on majority poll and other common rules. Finally , computation results are discussed. The proposed research method in this paper is very simple and satisfactory, all of these are very valuable to engineering applications of biometrics.

中图分类号: 

• TP391.4