| 基于自适应特征选择和SVM的图像分类的研究 |
| |
| 引用本文： | 潘崇,朱红斌.基于自适应特征选择和SVM的图像分类的研究[J].计算机应用与软件,2010,27(1):244-246. |
| |
| 作者姓名： | 潘崇 朱红斌 |
| |
| 作者单位： | 丽水学院计算机与信息工程学院,浙江,丽水,323000 |
| |
| 摘 要： | 提出多特征结合的图像分类方法,分别提取颜色特征和LBP纹理特征,同时提出Adaboost算法对特征进行选择,选择最能表示图像的特征,这样既降低了特征的维数,又提高了分类的精度。最后对基于SVM的多类图像分类方法进行了研究,提出在二类支持向量机的基础上构造多类分类器的方法,实验结果表明,提出的方法能够很好地用于图像分类。
|
| 关 键 词： | 图像分类 支持向量机 特征选择 Adaboost算法 |
ON IMAGE CLASSIFICATION BASED ON ADAPTIVE FEATURE SELECTION AND SUPPORT VECTOR MACHINE |
| |
| Pan Chong,hu Hongbin.ON IMAGE CLASSIFICATION BASED ON ADAPTIVE FEATURE SELECTION AND SUPPORT VECTOR MACHINE[J].Computer Applications and Software,2010,27(1):244-246. |
| |
| Authors: | Pan Chong hu Hongbin |
| |
| Affiliation: | College of Computer and Information Engineering/a>;Lishui University/a>;Lishui 323000/a>;Zhejiang/a>;China |
| |
| Abstract: | An image classification method of multi-feature mergence is proposed in this paper,it extracts respectively the colour feature and LBP texture feature,meanwhile Adaboost algorithm is put forward for feature selecting.By choosing the feature which represents image the best,it can both reduce feature's dimension and improve classification precision.In the end of this paper,the method of multi-class image classification based on support vector machine is studied,and the approach of constructing the multi-class... |
| |
| Keywords: | Image classification Support vector machine Feature selection Adaboost algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
