摘要
提出了一种模糊统计方法的脱机手写体汉字特征提取方法,结合小波网格方法和汉字笔画密度特征方法对汉字进行特征提取,并运用支持向量机方法,通过机器学习对脱机手写汉字识别。仿真实验表明,支持向量机方法在脱机手写汉字识别中有良好的识别性能及模糊统计方法是有效的。
In this paper, a new feature extraction method based on fuzzy statistic feature was proposed. SUM,The theory of small - sample statistical learning proposed by Vapnik,was used for offline handwritten Chinese characters recognition. The feature date was extracted by three methods they are the density of Chinese characters stokes, wavelet transform and elastic meshing, and fuzzy statistic feature. The result of recognition shows that the SUM method can be used practically in off - line handwritten Chinese characters recognition and the new feature extraction method is effective and scientific.
出处
《计算机技术与发展》
2006年第10期104-107,共4页
Computer Technology and Development
关键词
支持向量机
脱机手写汉字体汉字
模糊统计特征
汉字识别
SVM
off- line handwritten Chinese characters
fuzzy statistic feature
Chinese characters recognition
