基于Gabor小波的维吾尔文笔迹的特征提取方法研究  

作　　者：阿布都热衣木.热西丁 

机构地区：[1]和田师专,新疆和田848000

出　　处：《和田师范专科学校学报》2010年第5期184-185,共2页Journal of Hotan Normal College

摘　　要：本文提出使用Gabor小滤提取维吾尔文笔迹的纹理特征方法。首先,通过预处理形成维吾尔文纹理图像。按照维吾尔文笔迹的特点,对Gabor滤波器进行改进。然后用一组Gabor滤波,过滤后获得图像的高维特征向量。最后,最邻近分类器(NN)分类器对不同17名人的维吾尔文笔迹样本进行分类实验,并获得79.8%的鉴别率。This paper proposes a method for tex^e feature extraction by Gabor wavelet filters for Uyghur Imadwriting image. The texture image is firstly formed by pre-processing steps. Gabor filter improved according to the characteristics of Uygher hand writing. Then higher dimensional feature vectors are constructed from the filtered texture images. Finally, experiments were performed using NN classifier to Uyghur handwriting samples from 17 different people and promising results of 79.8% cocrect identification rate were achieved.

关 键 词：笔迹鉴别 维吾尔文笔迹 Gabor小波NN 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]