遗传算法优化特征权值的支持向量机图像分类  

Iamges Classification Based on Genetic Algorithm Optimizing Features Weight and Support Vector Machine

作　　者：王东霞[1] 周观民[1] 

机构地区：[1]济源职业技术学院信息工程系,河南济源459000

出　　处：《电视技术》2015年第2期75-79,共5页Video Engineering

基　　金：河南省科技攻关项目(132102210229)

摘　　要：为了提高图像分类的准确率,提出了一种遗传算法优化特征权值的支持向量机图像分类(GA-SVM)。首先分别提取图像的颜色和纹理特征,然后采用改进遗传算法确定特征权值,最后采用支持向量机建立图像分类器,并对corel图像库进行仿真测试。结果表明,相对于其他图像分类算法,GA-SVM提高了图像分类精度。In Order to improve the accuracy of the image classification, a support vector machine image classification （GA-SVM） is proposed based on genetic algorithm optimizing feature weight. Firstly, the color and texture of image are extracted. Then the feature weight are determined by the improved genetic algorithm. Finally, image classifier is established by support vector machine and the simulation test is carried out on core image library. The results show that the GA-SVM algorithm can improve the image classification accuracy compared to other image classification algorithm.

关 键 词：图像分类 特征加权 支持向量机 遗传算法 

分 类 号：TN911.73[电子电信—通信与信息系统]