Published Paper Details:

Offline signature verification is an effective biometric authentication method with widespread applications in areas including finance, law, and security. However, because of variations in the way that each person writes, as well as the existence of highly skilled forgeries, distinguishing real from false signatures can be very difficult. As a result, a number of researchers have proposed various methods for investigating the validity of signatures by applying principles of image processing and machine learning. Typically, researchers will preprocess the image of the signature, extract features related to it, and then classify these features to determine if a signature is either real or false. Historically, researchers have used many different methods to accomplish signature verification, including Support Vector Machines, Artificial Neural Networks, Hidden Markov Models, and Dynamic Time Warping, as well as more recently developing sophisticated detection models using deep learning architectures like Convolutional Neural Networks to improve accuracy. The purpose of this paper is to provide a survey of the current methods for offline signature verification to detail the advantages and disadvantages of each method. In addition, by comparing the strengths and weaknesses of their methodologies, researchers and practitioners can gain insight into what types of improvements can be made to signature verification systems in order to enhance their accuracy and dependability.