Abstract: In view of the demand on handwriting Uyghur letter on handset, according to the limitation of handset hardware resources, the Uyghur written language character is divided into two parts. The main body stroke, the attached stroke and so on, totally 21 characteristics as the characteristic set, build the Uyghur written language character recognition data sample template storehouse. Gathers current input coordinate sequences, classifies sketchily on the stroke number, respectively extracts the main stroke characteristics and the appendix characteristics and indicates with the binarization data. Carries on the XOR operation with the binarization and the character sample in template storehouse, and takes the sample which has least 1 integer in result after operation as candidate character. An average correct recognition rate achieves 97%.

Key words: Uyghur, mobile phone, characteristic code, handwriting

摘要： 针对手机手写维吾尔文字母输入需求，结合手机软硬件资源局限性，提出将维吾尔文字符分成主体笔画与附属笔画两部分。采用笔画数、主体笔画结构特征、附件结构特征等21个特征为特征集，建立维吾尔文字符识别数据样本库。按笔画数粗分类，分别抽取主笔画的特征与附件的特征并用二值化数据表示，与样本库中特征数据进行异或运算，取运算后1的个数最少的样本为识别样本并给出候选样本，平均一次正确识别率达到97%。

关键词: 维吾尔文, 手机, 特征编码, 手写

CLC Number: 

• TP391.43