基于广义Hough变换的手写文档整词定位

Word Matching on Handwritten Document Based on GHT

本文研究了在手写文档中定位目标整词的应用问题。将广义Hough变换的应用范围从图形匹配拓展到文档图像中的单词匹配。利用SIFT算法提取边界特征,将广义Hough变换的投票结果归一化到灰度值范围,形象地给出了单词位置检测的结果。最后将该方法应用到维吾尔文手写文档的整词定位问题中,取得了良好的效果。

This paper presents the study of the target word matching on handwritten document. Generalize Hough Transform （GHT） is introduced into the application of handwritten document analysis. The edge features are extracted by SIFF, and then the voting result of GHT is nomalized into gray values, which help us understand the localization of target word in the document. Finally, the proposed method is applied in word matching on handwritten document in Weiwur language.