基于Gabor函数的汉字字体识别

Chinese Character Font Recognition Based on Gabor Function

汉字字体识别是当前中文OCR系统需进一步解决的问题之一 ,字体信息属于汉字笔画上的微细结构部分 ,抗干扰能力差 ,这样基于局部微细结构分析的方法是具有较大难度的 .这里采用了一种新的基于Gabor函数的全局纹理分析的识别方法 .在实验中选用了汉字 4种常用标准印刷字体 (宋、仿宋、黑、楷 )的样张 ,以证实这种方法的有效性 .

Chinese character font recognition is one of the problems that needs more processing in current Chinese OCR systems .The font information belongs to the microstructures in Chinese character strokes, so it has a weak capacity of anti-jamming.Therefore,the recognition method based upon local microstructures analysis prossesses biggish difficulties. In this paper ,authors adopt a new global texture analysis based approach using Gabor function towards Chinese character font recognition, and choose four frequently used Chinese character standard printed font samples(Song ,Fangsong ,Hei,Kai) to demonstrate the potential of the technique.