核函数FISHER鉴别在性别鉴别中的应用

Applications of Kernel Fisher Discriminant on Gender Classification

该文在线性FISHER鉴别的基础上,探讨了核函数FISHER鉴别原理;然后将核函数FISHER鉴别用于性别鉴别中,并提出了相应算法;对ORL人脸图像数据库进行实验;详细分析了算法在不同样本分布和不同参数条件下的识别结果;表明了在两类模式识别中核函数FISHER鉴别理论具有良好的识别效果及极佳的推广能力。

In this paper,Kernel Fisher Discriminant (KFD)is discussed on the basis of Linear Fisher Discriminant (LFD).Corresponding algorithm is proposed,applied in gender classification,and analyzed in detail under the conditions of different sample distribution and different parameters.Experimental results on ORL face database demonstrate that KFD embraces good recognition effect and strong generalization capability in solving two-class pattern recognition problems.