基于连通域的字符孔洞提取算法及其实现

An Algorithm for Character Hole Extraction Based on Connected Components and Its Implementation

提出了基于等价类行程的连通域算法,通过计算字符图像背景区域连通域提取手写字符图像的孔洞特征。为将相关行程合并成连通域,构造一个等价类数组记录和传递相关行程间的连通关系。手写数字字符图像的实验结果表明,本文算法能准确有效地提取手写数字字符孔洞特征,可用于字符分类识别。

This paper presents an algorithm of connected components based on combination of equivalent run-lengths.Through calculating the connected components of the background area of character image,hole features of a handwritten character image can be extracted.In order to combine correlative run-lengths with a connected component,an equivalent array is constructed to record and transfer the relationship of the connectivity of run-lengths.The results obtained from the test for handwritten numeral image show that the algorithm can capture the features of handwritten numeral characters accurately and efficiently and this method can be used to classify characters.