节点文献
颜色和形状特征的商品图像检索技术的应用
Application of a Product Image Retrieval Based on Color and Shape Feature
【摘要】 针对基于内容的商品图像检索,提出了以颜色和区域形状为特征的检索技术方案.以非等间隔量化方法提取颜色直方图,同时设置显著水平阈值提取视觉主色,从而获取颜色特征.利用二维函数矩构造出的Hu矩特征值,并对矩特征值进行权重设置,提取了比较符合人类视觉的区域形状特征.两种特征组合检索,实现了商品图像的较高相似性度量的匹配.
【Abstract】 For product image retrieval,the method of retrieval based on color feature and shape feature is proposed.To obtain color featrue,Non-interval quantitative measure is used to extract color histogram,and setting a value number is to find the primary color.Implementing the algorithm of two-demensional function moment that is weighted to get the shape feature.As these two features are combined in retrieval,the results of math in product image achieve a highly level.
【关键词】 图像检索;
颜色特征;
形状特征;
颜色直方图;
二维函数矩;
相似性度量;
【Key words】 image retrieval; color feature; shape feature; color histogram; two-demensional function moment; similarity measure;
【Key words】 image retrieval; color feature; shape feature; color histogram; two-demensional function moment; similarity measure;
- 【文献出处】 计算机系统应用 ,Computer Systems & Applications , 编辑部邮箱 ,2012年09期
- 【分类号】TP391.41
- 【被引频次】4
- 【下载频次】113