16-03-2017, 04:29 PM
There are a number of biometric methods today, p. Signatures, fingerprints, Iris, etc. There is considerable interest in authentication based on the handwriting signature verification system as it is the cheapest way to authenticate the person. Fingerprints and iris verification require the installation of expensive equipment and therefore can not be used in day-to-day places like banks etc. As forensic experts can not be employed in each place, there has been considerable effort towards the development of algorithms that could verify and authenticate the identity of the individual. Signatures are often not even readable by humans. Therefore, a signature is treated as an image that carries a certain pattern of pixels belonging to a specific individual. The problem of signature verification, therefore, refers to determining whether a particular signature really belongs to a person or not.
Signatures are a special case of handwriting in which special characters and flourishes are viable. Verification of the signature is a difficult problem of pattern recognition, since no two genuine signatures of a person are precisely the same. Its difficulty also stems from the fact that qualified counterfeits follow the genuine pattern as opposed to fingerprints or irises where the fingerprints or irises of two different people vary widely. Ideally, interpersonal variations would be much more than intrapersonal variations. Therefore, it is very important to identify and extract those characteristics that minimize intrapersonal variation and maximize interpersonal variations. There are two approaches to signature verification, both online and off-line, differentiated by the way data are acquired. In case of connection, the signature is obtained on a piece of paper and analyzed later. While in case of online signing you get it in an electronic tablet and pen. Obviously, dynamic information such as speed, pressure is lost in case of connection unlike online case.
