SVM在小字符集脱机手写体汉字识别中的应用研究  

Application Research of SVM for Small-set Off-line Handwritten Chinese Characters Recognition

作　　者：石繁槐[1] 童学锋[1] 

机构地区：[1]同济大学计算机科学与工程系,上海200092

出　　处：《计算机工程》2002年第6期154-155,189,共3页Computer Engineering

摘　　要：提出将方法引入小字符集脱机手写体汉字识别,利用较少的训练样本就可以达到比较理想的识别效果,并在小样本学习的情况SVM下同最小距离法和多重相似度法的识别率和识别时间作了比较,说明方法在小字符集脱机手写体汉字识别中的实用性。This paper applies the SVM method to small-set off-line handwritten Chinese characters recognition and acquires satisfying effect of recognition by using small number of training samples. At the same time, the comparison between its recognition rate and recognition time and those of minimal distance method and multiple similarity method under small-sample training shows that SVM method can be used practically in small-set off-line handwritten Chinese characters recognition. ;;

关 键 词：SVM 小字符集 脱机手写体汉字 汉字识别 模式识别 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]