一种基于形状特征的图像检索方法  

AN IMAGE RETRIEVAL METHOD BASED ON SHAPE FEATURE

作　　者：蔡菲[1] 蔡珣[2] 史同广[1] 王京卫[1] 

机构地区：[1]山东建筑工程学院GIS教研室,山东济南250014 [2]山东大学计算机科学与技术学院,山东济南250061

出　　处：《计算机应用与软件》2005年第12期98-99,共2页Computer Applications and Software

摘　　要：本文主要提出了一个在形状轮廓的傅立叶描述子基础上,进行基于形状特征的图像检索的理论方法。该方法首先对图像进行预处理,利用多阈值颜色分割法提取出图像的前景对象,再采用轮廓跟踪法得到该对象的形状轮廓特征,并采用傅立叶描述子对轮廓加以描述。最后利用欧几里得距离进行相似性比较。为了验证方法的有效性,本文建立了一个基于形状检索系统的框架,实现结果和算法分析表明,该理论方法的算法复杂度为O(n2logn),对基于形状检索的图像检索系统是有效和合理的。A new method for CBIR based on shape Fourier descriptor is presented. The first step of this method is pre-processing the image ,then extracting the foreground object using multl-threshold image segmentation based on color feature, and then use contour tracking meth- od to obtain the contour feature which later described by Fourier descriptor using FFT, at last, using Euclid distance to measure the similarity between the original picture and target picture, In order to verify the effectiveness of the method,we build a CBIR prototype and also describe the framework for it, By several experiments,the results show that this method is effective and also by algorithm analysis,we conclude that the method is rational for image retrieve as the compute complexity is 0（n^2logn）,

关 键 词：图像检索 形状 傅立叶描述子 图像预处理 多闽值颜色分割法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP391[自动化与计算机技术—计算机科学与技术]