机构地区: 华南理工大学
出 处: 《华南理工大学学报(自然科学版)》 1998年第9期73-77,共5页
摘 要: 提出一种新的汉字结构特征提取方法.利用二维小波变换对汉字图像进行分析和处理,分别在其高频子图像中提取汉字的横、竖、撇和捺信息,使结构特征的提取比用其它方法更简单、明晰. A new method for structure feature extraction in the recognization of handwriting Chinese character is presented in this paper. By using wavelet transformation the image of Chinese character is analyzed and proceeded afterwards. Then the horizental, vertical, left_falling and right_falling strokes of character are extracted from the high frequence wavelet subimages separately. The method is greatly simplified and distinct as compared with other methods.
领 域: [自动化与计算机技术] [自动化与计算机技术]