Font Size: a A A

The Research On Handwritten Signature Verification

Posted on:2009-11-22Degree:MasterType:Thesis
Country:ChinaCandidate:M YangFull Text:PDF
GTID:2298360245488824Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Due to the popularity of computer and Internet, communications among people become more and more frequent. Despite of the convenience, it comes with some security problems. Traditional methods for recognition and verification fail to satisfy the society demands. Fortunately, development of the techniques for recognition and verification of identification based on biometric features provides a more convenient and more reliable solution.There are various kinds of biometrics that can be utilized for personal identification. With the popularity of credit cards, signature plays a very important role in authenticity and authorization. People are getting familiar with the process of handwritten signature verification. The automatic signature verification could be off-line or on-line. Comparatively speaking, the former is more adaptive in equipment involvement and can be applied in more fields, but more difficult to manipulate due to the loss of dynamic writing information. In this thesis, off-line Chinese signature verification is studied. The main research work is as follows:1. Feature extraction based on contour pursuit. It attempts to use the coordinate data of the traced closed-contour as bases to extract features. As the closed-contour can trustily reflect structural information of original signature. The proposed system starts with a closed-contour pursuit algorithm. Marr wavelet is selected as mother wavelet in this paper. The coordinate data of the traced closed-contour are decomposed into multi-resolution signals using wavelet transforms. Then the fractal dimension corresponding to the coordinate curves are acquired as features for verification.2. Gridding feature extraction based on Contourlet. Contourlet not only inherits the main traits of wavelet transform, such as multi-scale and time-frequency information, but can also capture direction characteristics. It can hold the geometrical structure of images, and implement a true sparse representation. Handwritten signature possesses abundant direction characteristics, thus Contourlet transform (CT) can seize the structural features of images effectively, which is in favor of the correct identification. In this paper, structural feature and statistical feature, two kinds of traditional features, are combined by virtue of grid thought. After dimensionality reduction to extracted eigenvector by Karhunen-Loeve (K-L) transform, genuine signatures and forgeries are distinguished through support vector machines (SVM). In order to improve Contourlet gridding feature, we apply fraction dimension in it. And then self-similarity properties of sub-blocks are captured, a better verification effect is reached.
Keywords/Search Tags:Handwritten Signature Verification (HSV), Feature Extraction, Contour Pursuit, Contourlet Transform (CT)
PDF Full Text Request
Related items