Online Signature Verification: A Review
Jawad-ur-Rehman Chughtai, Shehzad Khalid, Imran Siddiqi · 2014
From the last few decades, online signature verification (OSV) has become a hot research topic and have been employed in many application areas such as banking, law enforcement, industry, and security control etc. The growing security needs of today's society exert a pull on researchers to work in this area. A number of techniques along with their variations have been proposed in the realization of a fool proof & reliable signature verification system. Dynamic Time Warping (DTW), Hidden Markov Model (HMM), Support Vector Machine (SVM) and Neural Networks (NN) are the most promising approaches amongst the others. In this paper, we have presented a review of research carried out in recent past in the field of online signature verification and made a qualitative analysis of these state-of-the-art approaches. Today's society demands secure means for person's authentication. Traditional authentication methods are based on the knowledge (password, Personal Identification Number) or on the possession of a token (Identification card, keys), which can be forgotten or stolen. This fact makes biometrics to take its place as an alternative method for person's authentication and identification. Besides many other verification methods like fingerprints, iris, etc. Signature verification, a behavioral trait is one of the promising way to authenticate a person's identity. This paper is focus on the qualitative study about the signature verification techniques. The term biometrics refers to an individual's recognition based on personal distinctive characteristics. Two types of biometrics can be defined by taking into account the personal traits which are physical or behavioral. The physical are about catering the biological traits of users, for instance, fingerprint, face, hand geometry, retina, and iris. The latter takes into account the behavioral traits of users, such as voice or handwritten signature. Biometric system is an advanced method to induce security and is mainly employed for personal authentication. Handwritten signature comes into sight as the most socially undertaken and renowned method for individual verification among all other existing biometric authentication systems (14). A signature is a handwritten depiction of someone's name or some other mark of identification that a person writes on documents or a device as a proof of identification. The formation of signature varies from person to person or even from the same person due to physical & mental condition at that time, geographical location, age and other factors. The primary focus of a signature verification system is the detection of forged and imitated signatures (variations) generated by imposters, for instance, unskilled and skilled forgery. The intention behind signature verification systems is to minimize the false acceptance rate (FAR) and false rejection rate (FRR) but the two terms are inversely proportional. Signature verification can be viewed as offline or static signature verification & online or dynamic signature verification from data acquisition standpoint. In offline signature verification, signatures are recorded as images on paper which can later on be transformed into computer by means of a scanner and processed using offline verification stages. Offline signature verification is carried out on static features like shape, style variation, distortion, rotation variation, etc. on the other hand, Online signature verification makes use of dynamic features e.g. pressure, velocity ,stroke length, pen up/down time, etc. along with the shape of the signature (12).One of the key requirement of a verification system i.e. accuracy, can be achieved with greater precision due to the availability of dynamic information in online signature verification system as compared to offline signature verification systems. OSV is accepted far and wide by the communities for verification purposes as its more secure method than already available methods in use. It's difficult for imposters to copy all attributes (speed, pressure) along with the shape as it's present in the genuine signature (7). Due to the increasing popularity of the input capturing devices e.g. tablets, PDA's