层次型支持向量机人脸检测器  

Face detection based on hierarchical SVM

作　　者：陈超[1] 宋宇[1] 任德昊[1] 陈果[1] 

机构地区：[1]成都信息工程学院通信工程系,四川成都610225

出　　处：《中国测试技术》2007年第2期61-63,111,共4页CHINA MEASUREMENT & TESTING TECHNOLOGY

摘　　要：大量的数值实验表明,在非线性支持向量机中,核函数的选取对支持向量机性能的影响很大。核函数的选择一直是一个难题。本文利用遗传算法的全局优化能力比较并分析了多项式核函数和高斯核函数的检测正确率和支持向量个数两项指标,提出了一种层次型支持向量机人脸检测器。实验结果证明,本文的方法确实取得了较好的效果。Lots of experiments indicated that the selection of kernel function is very important to non-linear SVM. On the other hand, it＇s difficulty to select a appropriate kernel function. In this paper, genetic algorithm based kernel parameter optimization method was applied to face detection. Based on this method, a hierarchical SVM classifier was provided and realized. The experimental results have shown that the proposed method can locate faces accurately from the input image with complex background.

关 键 词：向量机 核函数 遗传算法 

分 类 号：TM930.12[电气工程—电力电子与电力传动]