基于SVM的多特征手写体汉字识别技术

The Technology of Multiple Features Handwritten Chinese Character Recognition Based on SVM

针对传统的模板匹配法对汉字的识别率较低,文中提出一种基于SVM的多特征手写体汉字识别技术。在提取网格特征的基础上增加对汉字质心特征、笔划特征、特征点的提取,并采用SVM算法构造分类器,实现对手写体汉字的识别。实验结果表明,该方法的平均识别率为95.9%,高于传统的模板匹配法。

To solve the recognition rate of traditional template matching method is not high for Chinese charac- ter, a new method of multiple featureshandwritten Chinese character recognition based on SVM is proposed. In addition to the extraction grid features, also extract the centroid feature, stroke feature, feature point, and use SVM algo- rithmeonstruetclassifierto achieve the recognition of handwritten Chinese characters. Experimental results show that the average recognition rate of the proposed method is 95.9% higher than that of the traditional template matching method.