Offline signature verification system using hidden markov model (HMM) analysis of varying number of states and state transition topology

• Main Authors: Bakri N.B., Ahmad S.M.S., Shakil A.
• Other Authors: 36805728300
• Format: Article
• Published: 2023
• Subjects: Hidden Markov; Offline Signature Biometrics; Biometrics; Image processing; Design and Development; Forged signature sample; Hidden Markov modeling; Number of state; Off-line signature verification; Signature images; State transitions; Testing results; Hidden Markov models

This paper describes the design and development of an offline signature verification system that is based on Hidden Markov Modeling (HMM) technique performed on a series of a localized direction feature extracted from a scanned signature image. It also describes the analysis of the testing results by varying the number of HMM states (5, 6, 7, 8, 9 and 10 respectively) and their state transition topology. The testing reported in this paper has been carried out on signature samples of 100 users which contain both their genuine as well as their skilled and random forged signature samples counterparts. The chosen algorithm is simple to be implemented which results in fast verification operation, and at thesame time is reliable in detecting forgeries. � 2009 WASET.ORG.