一种新的联机签名鉴别方法

A New Method of On-line Signature Verification

联机签名技术是基于人体特征的身份鉴别领域的重要技术之一。该文采用改进的方向链码方法对签名特征作了十分有效的提取,并应用支持向量机(SVM)模型进行身份鉴别。实验结果显示,在只有少量真样本进行训练的情况下,对于非专业模仿签名取得了错误拒绝率(FFR)2.22%,错误接受率(FAR)1.91%的较好鉴别效果。这充分表明这种方法具有很好的使用和推广价值。

On-line Signature Verification is an important technique in the research field of individual verification and identification based on human biological and behavioral characteristics.This paper propose s a new mended directional chain-code method to extract effectively signature feature,then use SVM(Support Vectors Machine)method to classify signatures.The experiments show this method can obtain good signature verification performance that FFR decreases to2.22%and FAR decrease to1.91%for no-special imitating signatures when having only little true samples to train model.The research indicates the proposed method is worth being adopted widely.