在许多文字识别系统中, 字符切分是预处理阶段的一部分, 其目的是从文本图象中分离出字母图象。而后才能针对切分后的每个字母进行识别。在具有连体特征的文字中, 字符切分就显得特别重要, 因为字符切分的准确与否直接影响字符的识别。维吾尔文就具有这种明显的连体特点, 本文主要讨论了采用抽取投影特征的方法, 实现了多字体维吾尔文的行切分、字切分和字符切分。

In many OCR systems , character segmentation is a necessary phase for character recognition. it is very important and difficult to segment characters in cursive script characters , because the incorrect segmentation affects the result of the characters recognition. Uygur characters are featured with cursive script . In this paper , we present a method of segmenting Uygur printed characters that is based on the projection of character image , to realize segmentation of line 、word and characters from a scanned image page.