一种新的基于颜色特征的图像检索方法

New Color Based Image Retrieval Method

提出一种新的基于颜色特征的彩色图像检索算法.该方法首先利用颜色的距离矩阵,对样本图像进行非监督的颜色量化,得到调色板,然后将待检索图像按最小距离映射到该调色板中,这样对于所有图像,就可以得到基于样本图像调色板信息的统一的颜色直方图.这种颜色量化算法精度高于有监督的量化,而速度又明显高于传统的无监督的颜色聚类.此外,结合颜色在图像中的散布情况,综合颜色的统计特征与空间分布特征来描述图像内容,保证图像内容描述的精确性.实验证明,利用该算法的检索效率较高,检索结果也能够较好的满足人的视觉感受.

A new color-based image retrieval method is proposed in the paper.First,according to the distance-matrix of color,the sample image is clustered in a self-organizing way,thus its palette can be obtained.Then the other images in the database can be mapped into the palette in terms of min-distance.In this way,all the images can have a uniform histogram based on the same palette.The quantization precision of this algorithm is higher than that of supervised method and its efficiency well than unsupervised way.Finally,the distributing conditions of colors are considered simultaneously.Therefore the content of image can be described more accurately since the statistic and special features of color are both used.The experiments show that the method mentioned above has high efficiency and its retrieval results match human visual percept well.