preview

On-line Signature Verification Based on Forward and Backward Variances of Signature

Powerful Essays
Signature is one of behavioral characteristics of human that can be used to identify individual. Widely implemented in everyday life, signature has very high social acceptance cite{jain1}. Signature mostly has been used to verify important documents, such as bank checks. However, in many occasions, human eyes cannot distinguish genuine from forgery signatures, that will become a serious problem in security system. Therefore, the necessity of high accuracy automatic signature verification systems is increasing nowadays.

Signature verification is divided into two classes based on the type of acquisition of signature , on-line and off-line signature verification cite{impedovo1}. On-line signature verification uses digital input device, such as digital tablet and digital pen, to acquire data during signing process in the form of sampled signals. On the other hand, off-line signature verification performs static data acquisition using scanner. In this case, the data is represented as images.

The fluctuation of handwriting, the limited number of training data and the difficulty of extracting stable feature of signatures are some problems that must be faced in signature verification. Many works in on-line signature verification have been done in order to find the accurate method cite{jain1,impedovo1,plamondon1}. Among those proposed methods, a lot of researchers have been used $x$ and $y$ positions of signatures directly. However we consider that it is not effective because genuine and forgery signatures have quite similar shape and they result the difficulty of the verification process. To overcome the problem, new signing features related to $x$ and $y$ positions of signature called forward and backward variances of signature are ...

... middle of paper ...

... Image and Signal Processing, vol. 150, no. 6, pp. 395–401, 2003.
ibitem{canggih1}
C. P. Wibowo, P. Thumwarin, and T. Matsuura, On-line Signature Verification Based on Angles of Pen-Motion. 32nd International Conference on Simulation Technology (JSST), 2013.
ibitem{taherzadeh1}
G. Taherzadeh, R. Karimi, et al. Evaluation of online signature verification features. 13th International Conference on Advanced Communication Technology (ICACT), 2011.
ibitem{nakanishi1}
I. Nakanishi, S. Nakatani, et al. Simply Partitioned DP Matching and Threshold Equalizing in DWT Domain On-line Signature Verification. International Conference on Frontiers in Handwriting Recognition (ICFHR), 2010.
ibitem{yu1}
Q. Yu, W. XingXing, et al. Learning Mahalanobis distance for DTW based online signature verification. IEEE International Conference on Information and Automation (ICIA), 2011.
Get Access