论文部分内容阅读
本文针对在线汉字签名鉴别系统中存在的误拒绝真实签名问题和误接收伪造签名问题,提出了相应的解决方法。首先引入伪匹配起点的概念,使得用动态规划算法求出的最优匹配路径更具合理性;采用了多模板的方法,认为最终的鉴别距离是被测签名与多个参考签名距离的平均值;根据最小风险贝叶斯决策规则,求出距离阈值h使系统的平均风险达到最小。实验结果证明了本文提出方法的有效性。
In this paper, aiming at the problems of false true signature and wrongly received forgery signature in online Chinese signature authentication system, the corresponding solutions are proposed. Firstly, the concept of pseudo-matching starting point is introduced to make the optimal matching path obtained by dynamic programming algorithm more reasonable. The multi-template method is adopted to determine the final identification distance, which is the average distance between measured signature and multiple reference signatures. According to the least-risk Bayes decision rule, the distance threshold h is calculated to minimize the average risk of the system. Experimental results show the effectiveness of the proposed method.