基于模糊判别准则和特征加权模板匹配的字符识别

Character Recognition Based on Fuzzy Identification Criterion Feature Weight Template Matching

字符识别算法的设计作为光学字符识别的最关键步骤,直接影响到字符识别的精度和速度。运用特征加权模板匹配算法并结合基于模糊原理的判别准则,设计了一个字符识别算法。该算法从当前应用最广泛的模板匹配法入手,对标准模板匹配算法加以改进。在此基础上利用模糊理论作为判断标准来进行字符识别。所提算法能够充分区分开字符笔画和非笔画部分对字符识别的影响大小,从而有效地提升了识别率和鲁棒性。从实验结果来看,该算法性能较优。

As the most key step of optical character recognition, design of character recognition algorithm directly affects the precision and the speed of character recognition. Combined feature weight template matching algorithm with fuzzy identification criterion, a character recognition method was designed. Based on template matching algo rithm which is widely used, the designed method firstly improved the standard template matching algorithm and em ployed fuzzy theory as identification criterion to recognize character. The proposed method can fully separate the influ ence on character recognition which was caused character stroke and character tailing, so as to effectively improve the recognition rate and robustness. The experiment results show the designed method has better performance.