基于第二代Bandelet的脱线签名识别

Handwritten Signature Verification Based on Second Generation Bandelet

由于脱线签名鉴定丢失了在书写过程中的动态信息,鉴定难度大。本文针对脱线签名识别的特点,提出了基于Ban-delet变换的特征提取方式,将传统的结构特征和统计特征有效地结合起来。通过K—L变换降低特征向量的维数,然后采用支持向量机(SVM)的方法进行训练和识别。对400个手写样本进行了识别,实验证明该方法能有效提高脱线签名的识别率。

In this paper, we focus on off-line Handwritten Signature erification （HSV） and presents a new feature selection method based on Bandelet, which gives full play to the merits of both conventional structure feature and statistical feature. After dimensionality reduction to extracted eigenvector by K-L transform, support vector machines （SVM） is used to train and test the data. Experiments are made by 400 different handwritten signature stylebooks, The result of our experiment has confirmed the effectiveness of the proposed approach.