一种改进的基于颜色直方图的图像检索算法

An Improved Image Retrieval Algorithm Based on Color Histogram

使用颜色直方图进行图像检索是图像检索的常用方法,但是传统颜色直方图检索容易丢失空间信息,从而影响图像检索的查全率和查准率.针对这一问题,提出一种改进算法,该算法首先进行图像分块,然后通过对各分块颜色直方图加权的方法进行检索.实验证明,该算法可以提高图像检索结果的查全率和查准率.

The use of color histogram is a common method for image retrieval,since conventional color histogram retrieval method is prone to lose the spatial information of colors, thus affecting both the recall and the precision. A new image retrieval algorithm is proposed in this paper. According to the method,an image is cut into sub- blacks first;then the retrieval is performed by the weighted method of sub-block＇s color histogram. Experimental results show that both the recall and the precision are improved effectively.