基于颜色和空间特征的图像检索

在环形颜色空间划分和颜色视觉关注度的基础上,提出一种新的基于颜色和空间分布特征检索算法.首先,在HSV色彩空间下进行颜色量化并计算每种颜色的质心,进行环形颜色空间分割.然后计算各环形颜色空间直方图,在此基础上计算图像的空间颜色分布矩.同时计算各种颜色对应像素的平均视觉关注度,以此作为该颜色特征的视觉关注特征.最后对特征向量进行高斯归一化,采用特征向量的L1-norm距离计算彩色图像的相似度并进行图像检索.实验结果表明,本文算法比SCH及Geostat方法具有更好的检索效果.

A New Approach for Image Retrieval Based on Color and Spatial Features

A new kind of color image retrieval algorithm based on color and spatial features was presented.In the first place,annular color segmentation can be arrived according to center of mass of every kind of color.Then the spatial-color moment can be attained based on the annular color histogram.At the same time the average saliency map of all pixels of every kind of color can be vision attention of every kind of color.The spatial-color moments and average saliency map of every kind of color can be used as the character vector of color images.The Gaussian model is used to normalize the different sub-characters distance to the character vector.The similarity of the querying image and other images is computed by the L1-norm distance.Experiments indicate that this method give better performance than SCH and geostat.