基于广义Hough变换的手写汉字文档关键词提取

Keyword extraction from handwritten Chinese document image using improved generalized Hough transform

提出了一种基于广义Hough变换的手写汉字文档关键词提取技术。对于待提取的手写文档图像,采用字符像素逐点匹配和投票的方式进行广义Hough变换,在参数空间中定位出手写关键词图像的位置。本技术对传统的广义Hough变换进行了修改,突破了形状匹配需要完整轮廓信息的局限,简化了局部特征的计算,对手写汉字文档图像中具有局部形变、部分旋转和缩放的手写关键词能够有效提取。对于提取的相同关键词建立训练集,用签名识别的方法对书写者建模,能够达到书写者身份鉴别的目的。

In this paper, we present a keyword extraction methodology from handwritten Chinese document image using im- proved generalized Hough transform. In the voting phase, the features of each character pixel of handwritten character image are compared with the ones in the reference table. A vote is made in the parameter space when the features match each other and the location of the keyword is found by the cluster of the votes. In our method, the generalized Hough transform is modified so that the complete contour information of the shape is not necessary, and the computation of the local features is simplified. The hand- written keywords with local shape variation, slight rotation and zooming can be extracted from handwritten Chinese document image effectively. The same keywords from a same document can be used as training samples, which are used to model writers hased on the signature verification methods to accomplish the goal of writer identification.