| 基于神经网络和证据理论的离线签名识别 |
| |
| 引用本文： | 陈刚,李弼程,林琛.基于神经网络和证据理论的离线签名识别[J].微计算机信息,2005(21). |
| |
| 作者姓名： | 陈刚 李弼程 林琛 |
| |
| 作者单位： | 郑州解放军信息工程大学信息科学系
450002(陈刚，李弼程)，郑州解放军信息工程大学信息科学系 450002(林琛) |
| |
| 基金项目： | 河南省教育厅基金(编号:sp200303099)资助项目 |
| |
| 摘 要： | 论文提出了一种有效的基于神经网络和证据理论的离线签名识别方法。利用从签名图象中提取的三种特征分别构造BP神经网络分类器对签名图象进行初步识别,然后将各分类器的输出作为证据,利用证据理论融合来自不同分类器的输出得到最终识别结果。实验表明,该识别方法能有效提高离线签名的识别率。
|
| 关 键 词： | 离线签名识别 神经网络 证据理论 融合 |
Off-line Signature Recognition Based on Neural Networks and Evidence Theory |
| |
| Chen,Gang Li,Bicheng Lin,Chen.Off-line Signature Recognition Based on Neural Networks and Evidence Theory[J].Control & Automation,2005(21). |
| |
| Authors: | Chen Gang Li Bicheng Lin Chen |
| |
| Abstract: | An effective method based on neural networks and evidence theory for o ff-line signature recognition is proposed in the paper. Firstly, three sets of f eatures are extracted from a signature image, and then these features are fed to three BP neural networks classifiers for elementary recognition. Finally, recog nition result comes out by fusing the previous recognition results from different classifiers with evidence theory. Experimental results show that the proposed method is effective. |
| |
| Keywords: | off-line signature recognition neural networks evidence theory fusion |
| 本文献已被 CNKI 等数据库收录！ |
