对维吾尔文签名图像进行预处理(含灰度化、平滑、二值化、归一化、细化等)的基础上，提出基于密度特征的维吾尔文离线签名识别方法，结合签名样本的空间地位信息，对离线签名进行有效处理和识别。从维吾尔文签名样本库中选取75个签名者的(20个样本/人)1500个签名样本进行分类实验，得到最高为96%的识别率，实验结果表明，密度特征是维吾尔文手写签名的一种有效特征，能较全面地描述与捕捉维吾尔文手写签名的书写特点。%On the basis of preprocessing Uyghur signatures through gray scaling,noise removing,binarization,normalization and thinning etc.,a type of signature recognition method based on density features was proposed,and Uyghur signatures were effectively identified by combining the position information of signature samples.The classification experiments were carried out using 1500 Uyghur signatures selected from 75 persons (20 samples/person)in Uyghur handwritten signature database.96% of highest identification rates were obtained.Experimental results show that the density feature is a kind of effective feature for Uy-ghur handwritten signature,and it can comprehensively describe and capture the writing characteristics of Uyghur handwritten signature.
展开▼
