基于改进的加权Hausdorff距离的图像匹配  

Improved Approach to Image Matching Based on Weighted Hausdorff Distance

作　　者：蒋新土[1] 吕岳[1] 

机构地区：[1]华东师范大学计算机科学技术系,上海200062

出　　处：《计算机应用研究》2007年第4期182-183,238,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(60475006)

摘　　要：提出了一种改进的加权Hausdorff距离,并将其应用于字符图像的匹配。该方法根据字符图像的结构特征对字符不同区域设置不同的权重。实践表明该方法改善了图像匹配效果。An improved approach based on weighted Hausdorff distance with application to character image matching was proposed. Thanks to the idea that different feature should correspond to different weight, different weights were set for different regions of a character image in the proposed method. Experimental results showed that the improved method is promising and outperforms the original weighed Hausdorff distance.

关 键 词：相似性度量 HAUSDORFF距离 图像匹配 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]