现代电子技术Issue(15)：42-44,49,4.
基于分块颜色核特征的图像检索方法
Image retrieval method based on block color kernel feature
王琪 1杜娟 1徐国清 1程彬2
作者信息
- 1. 南阳理工学院 计算机与信息工程学院，河南 南阳 473000
- 2. 南阳理工学院 软件学院，河南 南阳 473000
- 折叠
摘要
Abstract
Color feature is the most intuitive features of images. Even though the traditional color histogram can describe re⁃spective color component in image,spatial location distribution information of respective color component is destroyed. It makes the image retrieval results inaccurate. Because of this reason,a method of block color kernel feature is proposed,which uses overlap block strategy to emphasize visual content of image center zone. Block color histogram is mapped to kernel feature space by the designed color kernel function,so the image color features are described more accurately. The experimental results show that in comparison with global color histogram method and block color feature method,the proposed method has higher precision ratio and recall ratio.关键词
颜色特征/图像检索/重叠分块/核函数Key words
color feature/image retrieval/overlap block/kernel function分类
信息技术与安全科学引用本文复制引用
王琪,杜娟,徐国清,程彬..基于分块颜色核特征的图像检索方法[J].现代电子技术,2015,(15):42-44,49,4.
