Abstract: This paper gives a recognition algorithm which combines 2 features of the hand-written Uyghur character. The algorithm extracts the real value Gabor energy feature and directional element feature of hand-written Uyghur character, combining the 128 dimension real value Gabor energy feature and the 128 dimension directional element feature together, uses KNN classifier to unitedly classify these two features. It recognizes the character feature and the handwriting feature for the samples in the database of the hand-written Uyghur character. The experimental result shows, the algorithm is more remarkable in recognition of the handwritten Uyghur character than using one feature. The algorithm can also be used in recognition of handwritten Arabic character.

Key words:  hand-written Uyghur character, Gabor filter, directional element, K-Nearest Neighbor（KNN） recognition classifier

摘要： 提出一种联合两种特征的手写体维文字符识别算法。该算法对手写体维文字符图像进行实值Gabor能量特征和方向线素网格特征的提取，将实值Gabor滤波器的128维能量特征和方向线素的128维网格特征结合起来，使用KNN分类器对两种特征进行联合分类。对手写体维文字符数据库中的样本分别进行手写体维文字符特征识别和维文字符笔迹特征识别。实验结果表明，和采用一种特征的识别算法比较，进一步提高了手写体维文字符的识别率。该算法也可用于手写体阿拉伯文字符的识别。

关键词: 手写体维文字符, Gabor滤波器, 方向线素, K最近邻（KNN）识别分类器