摘要
在基于内容的图像检索(CBIR)中,对于传统的颜色直方图完全没有空间分布信息,提出了一种新的分块划分,并且结合颜色特征的图像检索方法。该方法结合了图像的整体与分块颜色分布,两幅图像之间的相似度为整体相似度和分块局部相似度的加权和,并且在检索中加入相关反馈技术,针对检索结果适当地调整权值,以达到更新图像整体与局部颜色特征的权重的目的。最后,实验结果表明该算法能很好地提高检索性能。
In Content-Based Image Retrieval（ CBIR）, for the traditional color histogram no space distribution information, a new kind of block division is put fouward, and combined with characteristics of color image retrieval method, this method combining with the image of the whole and the block color distribution. In this paper, two image similarity between for overall similarity and blocking local similarity weighted, and in the retrieval join relevant feed- back technology, aiming at the retrieval results appropriate cumulative value, to update image whole and local color characteristics of the purpose of the weight. Through the final experimental results show that the new algorithm can improve the retrieval performance very well.
出处
《电视技术》
北大核心
2013年第7期29-32,共4页
Video Engineering
