基于分形特征的二值图像检索方法的研究

论述了基于分形特征的二值图像检索方法。图像的内容由4种特征来描述:图像分形维数、图像分形矢量、边界分形矢量和骨架分形矢量。实验表明该方法计算简单、有效,匹配快速,检索结果比较理想,只要查询图像在图库中,就一定能通过该方法检索出来。另外,该方法还具有一定的鲁棒性,证明这种检索方法是具有较大实用意义的。

Study of Binary Image Retrieval Method Based on Fractal Behavior

In this paper,a novel approach to feature extraction based on fractal theory is presented as a powerful technique in image retrieval.This paper presents a method that can be applied to extract the feature of binary images.The content of image is depicted by four features:fractal dimension of image,fractal vector of image,fractal vector of edge and fractal vector of skeleton.It is constructed by a hybrid feature extraction combining fractal theory,edge detection and skeletonization.Through extensive experiments with a database of2000trademark images,the method we demonstrated is robust and flexible and a high recognition rate is reported.