The verification of the signature is one of the most accepted biometric techniques, since a signature is a part of everyday life, although less accurate than biometric techniques such as the use of the iris. In this field, much attention has been paid to the characteristics, since a verification system must be able to overcome problems such as falsifications, insensitivity to intrapersonal variability and sensitivity to interpersonal variability. In this paper we present a simple and efficient approach to online signature verification based on a discrete cosine transform that has been applied to 44 time signals such as position, velocity, pressure and angle of the pen . The experiments are carried out in two reference databases, SVC2004 and SUSIG. The advanced feature selection algorithm is used to find subsets of features that work best. The proposed system was tested with different classifiers, with specialized falsification, and the same error rates were 3.61%, 2.04% and 1.49% for the SVC2004 Task1 & 2, Task2 and SUSIG databases, respectively .
Biometric verification techniques require the user to present identification information based on an unalterable personal characteristic. This may be a physical characteristic, such as a fingerprint or iris, or it may be a characteristic behavior, such as a signature or a voice. Signatures have been considered a typical form of authentication in our society for hundreds of years. Signature Verification is the most natural and friendly approach to personal authentication for many biometric-based verification systems.
A signature is a simple and concrete expression of the unique variations in the geometry of the human hand. It is known that the way a person signs his name is characteristic of that individual. Signatures are learned and acquired over a period of time rather than being a physiological characteristic, and are influenced by the physical and emotional conditions of a subject.
A signature verification system must be able to detect forgeries and, at the same time, reduce the rejection of genuine signatures. Signature analysis based on the method used to capture signatures is divided into two main categories: Offline and online. In offline verification, the signature patterns are signed on paper and then scanned using plate-type scanners. Online signature patterns have more information than offline patterns. Not only are there static geometric forms, but also dynamic writing information, such as velocity, acceleration and pressure, etc. Online signature verification methods have proven to be more accurate than offline methods.
Signatures are subject to intrapersonal variations. Therefore, a signature verification system is feasible only if the system is insensitive to intrapersonal variability, but sensitive to interpersonal variability. Even when they are insensitive to intrapersonal variations, the system must possess discriminatory power to thwart skilful counterfeiters.
Significant research has been carried out on the extraction and selection of features for the application of online signature verification. All of these features may be important for some problems, but for a given task, only a small subset of features is relevant. In addition to a reduction in storage requirements and computational cost, these can also lead to an improvement in overall performance. On the other hand, the selection of a subset of features requires a multi-criterion optimization function, eg. The number of characteristics and the accuracy of the classification.
