融合曲波变换和颜色直方图的图像检索

Novel image retrieval algorithm based on Curvelet transform and color histogram

研究了曲波变换和颜色直方图在图像检索系统中的应用,在此基础之上提出了一种融合Curvelet变换和颜色直方图的彩色图像检索的方法。该方法首先采用Curvelet变换提取彩色图像中的纹理特征,再利用颜色直方图提取颜色特征,最后分别用Manhattan距离和欧几里德距离进行纹理和颜色的相似性匹配。实验结果表明,融合曲波变换和颜色直方图的方法对图像检索非常有效。

Curvelet transform and color histogram used in image retrieval are studied in this paper.Moreover,a novel image retrieval algorithm which is based on Curvelet transform and color histogram is presented.Firstly,texture feature is extracted by curvelet transform.Then,color feature is represented by color histogram.Finally,image texture retrieval and image color match are respectively achieved by Manhattan distance and Euclid distance.Experimental results show that the novel image retrieval algorithm which is based on Curvelet transform and color histogram is greatly efficient for image retrieval.