基于SVM的静态手写签名识别方法研究

研究离线（静态）签名文字识别的问题.针对静态手写签名只有空间信息,没有时序信息的问题,提出了一种基于支持向量机的签名文字识别算法.算法首先通过图像预处理提取签名图像,然后利用防射变化对初始签名图像进行配准和矫正,并在此基础上利用滑动窗口在不同方向上提取特征.最终的识别器训练利用支持向量机完成.通过仿真实验证明,提出的新的方法较其他方法具有更高的识别准确率,可以有效的实现离线签名字体的识别,该技术可应用在个人身份认证和识别系统中.

A SVM Based Offline Signature Verification System

In this paper,we propose a new method for signature verification using local transform.The proposed method uses transform locally as feature extractor and Support Vector Machine(SVM) as classifier.The main idea of our method is using transform locally for feature extraction,against using it globally.The advantages of the proposed method are robustness to noise,size invariance and shift invariance.Having used a dataset of 750 signatures from 50 Persian writers,our system achieves good results.The experimenta...