基于着色处理的维吾尔文手写文本行分割

Coloring based handwritten Uyghur text line detection and separation algorithm

针对维吾尔文手写体文本中行分割问题,该文根据维吾尔文字包含区分字母的诸多点笔画和附加部分的特点,对文字主体笔画部分采用分段式前景涂抹算法,并将涂抹后图像的前景部分进行细化运算,通过细化线条与原图像的叠加、着色处理获取主体文本区域定位;然后通过重心辐射扫描的方法,解决了剩余细小笔画的文本行归附问题。实验结果表明:该文方法与常见的水平投影、分段式投影与基于涂抹方法相比具有更好的行分割效果。

The characteristics of Uyghur text which contains a large number of points and additional parts was taken into consideration to deal with the issues of text lines segmentation in Uyghur handwritten documents. Localization of main text lines was realized using the algorithm of piecewise foreground smearing followed by the foreground thinning operations, the overlapping of thinning lines with original images, and coloring process. The center of gravity radiation scanning method was developed to solve the belonging problems for remaining small strokes. Experimental results show that compared with the horizontal projection based, piecewise projection based, smearing based segmentation methods, the proposed method has better text line segmentation results.