基于分块主颜色匹配的图像检索

针对颜色直方图无法表达颜色空间分布信息的问题，提出一种基于分块主颜色匹配的图像检索算法，对图像进行新的重叠分块，提取各分块的主颜色，将简化的二次式距离计算方法应用到分块主颜色的匹配中，并对图像各个分块的颜色特征采用不同的加权系数。实验结果表明，该算法在保持较高检索精度的同时，能够降低特征空间维度，提高检索性能。

Image Retrieval Based on Sub-block Dominant Color Matching

In order to solve the problem of lacking the spatial knowledge of color histogram, an approach of image retrieval based on dominant color of partition is presented. It designs overlapped sub-blocks to get the color accumulative histogram in each block. Meanwhile, the reduced quadratic form distance method is applied to decrease arithmetic complexity. The similarity is calculated based on the accumulative histogram of each region. Experimental results show the algorithm can greatly reduce the dimension of color feature, and keep high retrieval precision simultaneity.