摘要
提出了一种新的相似性图像检索方法。首先分割图像,把一幅图像划分为若干子区域,抽取图像的分块颜色信息;再利用信息熵计算图像之间的颜色互信息。用这种方法对不同类型图像进行试验,结果表明,与其他颜色特征描述方法相比,该方法具有较好的效果。
A new content-based image retrieval method was proposed, Firstly, an image was segmented into several sub-areas, and the block color information was extracted from the image, Secondly, mutual information of color among images was computed using information entropy. Compared with the algorithms of other methods of color description, experiments with different kinds of images indicate that this approach has the advantages of good effect and high stability.
出处
《计算机应用研究》
CSCD
北大核心
2008年第1期281-282,共2页
Application Research of Computers
基金
国家自然科学基金资助项目(10661007)
江西省自然科学基金资助项目(0611009)
华东交通大学校立科研基金资助项目(06ZKJC03)
关键词
信息熵
图像分割
特征描述
互信息
information entropy
image segmentation
feature description
mutual information
