综合contourlet变换和颜色特征的图像检索算法

针对单一特征图像检索效率不高的缺点,提出一种基于轮廓波变换(contourlet)和颜色直方图的多特征检索方法.Con-tourlet变换采用"双重滤波器组"结构,具备多分辨率、多方向性和各向异性等性质,能以接近最优的方式表示图像的边缘.颜色直方图计算简单,具有旋转、平移的不变性.本算法充分利用Contourlet变换后系数分布特点和HSV空间颜色直方图的旋转平移不变性,构造图像的特征向量,以Euclidean distance作为相似度量标准进行图像检索.实验结果表明,提出的算法具有更高的查准率.

Content-based Image Retrieval Using Color Histogram and Contourlet Texture Feature

This paper proposed a content-based image retrieval method based on an efficient combination of color and texture features.Cumulative color histogram of the images in HSV color space is easy to compute,insensitive to rotation and scaling.Contourlet Transform(CT) has the characteristics of multiresolution、multi-directionality and anisotropy.This algorithm uses CT coefficients and cumulative color histogram of image to constrcte feature vector,Euclidean distance to measure similarity,finally realize feature matching and image retrieving.Experimental results demonstrate that the proposed method is effective and has higher precision ratio.