A Study About Stability Of On-line Handwritten Signature Based On Stroke Segmentation

The rapid development of the Internet has brought about great changes in people's work and life.People are taking advantage of the Internet to make work and life convenient when they adapt to the changes.With the emergence of e-commerce,information security has become a hot issue that people increasingly concern.However,the security of traditional identity certification based on identifying objects or the passwords of the users is not high.Because people can lose the objects or forget the passwords,which make the users feel uneasy and troublesome.At present,the identity certification based on biological characteristics has come into public attention.Biological recognition technology includes two categories mainly,one category identifies by physiological feature and another category identifies by behavioral feature.Biological recognition technology has characteristic advantages in the field of personal identification which include be able to take out,be not easy out and high recognition rate.On-line handwriting signature verification is a kind of biological recognition technology which is identified by behavioral feature.On-line handwriting signature is an unique biological certification technology with high practicability and simple operation.Moreover,its uniqueness and stability make it occupy an important position in the field of identity certification that is based on biological characteristics.Handwritten signature is common identity authentication in people daily life which operates easily and conveniently.The development of smartphones and tablet computers promotes the development of writing pad so that the price of writing pad gets more and more cheap and the performance of writing pad gets more and more reliable.The popularity of writing pad gets on-line handwriting signature more and more easy,provides equipment foundation for the development of on-line handwriting signature.On-line handwriting signature with high value in use and its practical significance have boosted the rapid growth of global e-commerce.Firstly,it connects the handwriting board to a computer to get data from the signature sample and saves them in format TEXT in the computer.Then the technology pretreats the acquisition data and removes interference and noise.This is the process of denoising aiming to reduce the workload in the later characteristics acquisition,which includes the treatment of the first stroke of a character,smoothness,uniformization,etc.Secondly,this paper analysis the strokes division of maximal angular velocity and minimum linear velocity,and confirms that the division results are different by experiment.This paper uses IMP value algorithm to strokes division,makes division points,starting points of signature strokes and terminal points of signature strokes to be special points of signature,divides the signature to small strokes segment.This paper judges the relationship of strokes segment by Euclidean Distance.This paper chooses four at least stable strokes segment to be judgement basis of signature in five signature samples.Finally,every volunteer writes ten real signatures,and then this paper chooses five signature samples randomly from all signatures.Compute the distance and the dissimilarity degree through dynamic time warping algorithm.Then set threshold based on statistics to design the classifier and achieve the effect of distinguishing the true from the false.Numerous experiments have proved that FAR can reach 3.1%and FRR 3.7%.As can be seen from the result,the research of this paper can effectively identify the signature certification.