分块检测与特征嵌入相结合的图像检索方法研究  

Research on Image Retrieval by the Combination of Block Detection and Features Embedment

作　　者：宁晶晶[1] 周海英[1] 

机构地区：[1]中北大学电子与计算机科学技术学院,太原030051

出　　处：《电脑开发与应用》2011年第6期24-26,共3页Computer Development & Applications

摘　　要：要想提高图像检索的效率,需要考虑到两个方面的问题,图像的分类和匹配,采用好的方法对图像分类,可以使图像检索的结果更精确,再结合好的匹配方法加快检索速度,就会使图像检索的效率大大提高。所描述的图像检索系统,在图像分类时,先分块检测,再将综合分块检测的结果对图像进行分类,取得良好的分类效果。在图像匹配过程中,引用水印作为索引,大大提高了检索的速度。实验证明,效果显著。In order to improve the efficiency of image retrieval,there are two aspects needs to consider,image classification and matching.With good method of image classification,the result of image retrieval will more accurate.Then combined with good matching method,will increase the efficiency of image retrieval greatly.This paper described a system of image retrieval,in classification aspects,integrate all the results of every block to classify them,in matching aspects,with the index of watermark,improve the retrieval speed greatly.Experiment shows that the effect is remarkable.

关 键 词：分块检测 图像匹配 水印 索引 分类 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]