基于颜色特征的图像检索技术

Color-Based Image Retrieval

在基于内容的图像检索(CBIR)中,颜色特征作为图像的一种重要视觉信息,由于计算简单,鲁棒性强,已经得到了广泛的应用。从图像的颜色特征出发,针对颜色空间(颜色空间的转换)、颜色的量化(HSV颜色空间)、颜色特征提取及表示和颜色特征的相似性度量做了系统讨论和研究。接着通过经典的分块颜色直方图算法来介绍基于颜色特征的图像检索全过程。最后总结了基于颜色特征的图像检索研究存在的问题以及进一步的研究方向。

The color features,as an important visual information of image,have been widely used in the Content-Based Image Retrieval（CBIR） because of the simple calculation and better robustness.According to the color features of the image,this paper makes systematic discussion and research on color space（color space transformation）,color quantification（HSV color space）,color features extraction and color characteristic similarity measurement.Then,through classic color segment histogram algorithm,the entire process of Color-based Image Retrieval is proposed.Finally,the existent problems and further research direction of the color-based image retrieval is summarized.