一种基于极值区域性特征的在线签名认证算法

Online Signature Verification Algorithm Based on Regional Feature of Extreme

通过对在线签名认证算法的比较分析,提出了一种新的算法来实现在线签名认证。该算法通过计算图像上点坐标(x,y)的导数值为零得到极值点,把极值用来作为分段边界,得到一组基于所定义极值特征的规则。阐述了提取签名中的区域性特征。实验表明,该算法的错误拒绝率和错误接受率都明显下降。

This paper proposes a new online signature verification algorithm based on the comparison and analysis of online signature verification algorithm. This algorithm is used to compute point coordinate（x, y） in pattern so that extreme point is acquired when derivatives of x and y are. Extreme is used as segmenting boundaries and a set of rules derived from the properties of extreme are defined. Regional features in signature are extracted and expanded fully. The experiments show that the false rejection rate and false acceptance rate are decreased obviously.