摘要
针对无限制手写体数字字符的识别问题进行了探讨．输入图像经过预处理后。进行特征抽取及分域处理．每一区域内对应一类基元、基元的类别由特征组合情况来确定．对骨架的拓扑结构进行全面描述．形成输入字符图像的一维链码．以此与参照链相匹配．根据匹配结果。进行识别．
A method for recognition of unconstrained handwritten character is probed. The feature extraction and the segmentation of input image are made after preprocessing. The primitive string is used to describe character skeleton and forms one-dimension string code of input character to match wtth the standard string. According to the resalt of matching. recognition is finished.
关键词
图像识别
数字字符图像
计算机
pattern recognition
feature
primitive
matching
