基于混合核函数的支持向量机在人脸识别中的应用研究

Face recognition based on multi-kernel support vector machine

针对支持向量机中的核函数选择和参数优化问题进行研究,结合局部性函数和全局性核函数的特点,形成由高斯核函数和多项式核函数构成的混合核函数,并运用于人脸识别,仿真实验结果证明了混合核函数的具有较高的识别率。

This paper researches the selection problem of kernel function and the parameters optimization problem Support Vector Machine （SVM）. Through the features of local kernel function and global kernel function, we mix the Gaussian kernel function and polynomial kernel function together and propose a new kernel function named multi-kernel function. Then we apply Multi-kernel function into face recognition and prove that multi-kernel function can achieve a higher recognition rate.