基于变精度粗糙集的脱机手写体汉字识别

Off-line handwritten Chinese character recognition based on variable precision rough set

文章为研究脱机手写体汉字识别问题,将汉字样本及其特征向量看作是一个信息系统,采用基于β近似依赖度的属性重要度定义作为启发式信息,设计出在变精度粗糙集模型下的特征属性近似约简算法,对手写体汉字信息系统中冗余特征属性进行约简,构建出脱机手写体汉字识别决策信息系统。识别过程中采用基于加权规则置信度的规则融合方法,进一步提高了脱机手写体汉字的可识别性和正确识别率。实验结果表明,该方法是有效可行的。

The Chinese character sample and its feature vector are considered as an information system to study the off-line handwritten Chinese character recognition.A heuristic algorithm of the attribute reduction in variable precision rough set is designed.The definition of attribute importance,which is based on β approximation dependence,is taken as heuristic information in the heuristic algorithm.The redundancy features of handwritten Chinese character are reduced by the heuristic algorithm to build up the off-line handwritten Chinese character recognition decision-making information system.Moreover,a kind of rules fusion method based on weighted rules confidence is proposed in the process of recognition to improve the identifiability and correct classification rate of handwritten Chinese character.The experiment results show that the method is feasible and effective.