在手写签名认证中,特殊点的抽取和匹配具有非常重要的意义,因为特殊点不仅可以作为签名的一种特征,还可以作为分段、签名对齐的依据.该文提出一种基于串匹配的方法对签名的特殊点序列进行匹配,提高了特殊点匹配的精度.同时通过计算特殊点类型所组成的字符串的编辑距离,能够很快地发现与真签名差距较大的自由伪造签名,从而提高整个系统的反应速度.
For on-line signature verification, extraction and matching of special points is one of the most important problems. Because the special points can be regarded as features, and used as references in segmentation and alignment. This paper proposed an approach based on approximate string match to match the sequence of special points. By calculating the edit distance between strings consisting of the types of special points, this method can quickly detect random forgeries. This can improve the speed of the system.