摘要
对无限制手写体数字的两大特点──畸变性和旋转性，提出了一种链匹配识别算法。全面描述了边界的拓扑结构（基元），根据形成输入字符图像的一维链码与参照链的匹配结果进行识别。压缩了原始图像信息，提高了处理速度，同时也达到了较高的识别率。
This paper puts forward a recognition algorithm for counter distortion and rotational variance of unconstrained handwritten numerals by string match.The primitive border is described thoroughly and one-dimension string code of input numerals is formed to match a standard string.According to the result of matching,recognition is one,so the original image data are reduced.The method is proved to be of high processing speed and recognition-rate.
