摘要： 提出一种基于改进的特征空间分离变换和支持向量机的人脸性别分类方法。在FERET人脸库和淮阴师范学院学生人脸库上进行实验，比较不同的特征提取方法和分类方法处理人脸性别分类问题的性能，结果表明，采用新方法在最优投影轴数和正确识别率方面均取得较好的结果，在2种人脸库上的正确识别率优于主成分分析方法和线性鉴别分析方法。

关键词: 主成分分析, 线性鉴别分析, 特征空间分离变换, 支持向量机

Abstract: This paper proposes a face gender classification method based on improved Eigenspace Separation Transform(EST) and SVM. Classification experiments are conducted on FERET database and student human face database of Huaiyin Teachers College to compare different feature extraction methods and classification methods of the human face on the issue of gender classification. The results show that the performance of the new method is satisfied, and it is superior to Principal Component Analysis(PCA) and Linear Discriminant Analysis(LDA).

Key words: Principal Component Analysis(PCA), Linear Discriminant Analysis(LDA), Eigenspace Separation Transform(EST), SVM

中图分类号: 

• TP391.41