摘要
细化处理能极大地消除光学文字、图像中的冗余信息量,对其存储、处理、识别、重建等非常重要.本文中我们提出比较符合人眼视觉效果的用于文字识别的快速自适应细化算法(ATA),它具有对称细化处理特点,细化的结果能很稳定地保持原始文字的结构信息,又给识别带来了方便.ATA算法在DEC-5000工作站上执行,每秒钟至少可细化375个手写数字(按占cup时间计算,矩阵大小38×38),细化的对象可以是印刷体或手写体的文字、字母、数字,也可以是工程线画图等.
A new method of more shape-adaptive thinning algorithm and its implementation is proposed. The resultant performances are presented also. It is mainly used for obtaining skeletons of binary handwritten character images after normalization. The bast property is preserving the original structure information of input characters to be recognized by OCR systems, which is specially helpful for the recognition of handwritten characters such as Chinese characters, English letters and numerals. Moreover, the thinning speed is higher than 375 characters per second on DEC-5000 Workstation. And it is more shape-adaptive in practical use. It is easier to implement the algorithm by means of hardware chips.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
1997年第2期140-146,共7页
Pattern Recognition and Artificial Intelligence
关键词
模式识别
文字识别
图像识别
自适应细化法
Thinning, Pattern Recognition, Handwritten Characters, Templates