Modified Shape Context for Signature Verification of Automated Cheque Authentication System?

Every person has a unique signature hence it is considered as one of the biometrics for the purpose of authentication. Signature is an indivisible part of any bank cheque but it can be copied by skill or by mere observation. To avoid such frauds many techniques have been found till date. Research here includes shape contexts for authentication of signature on bank cheque. Shape Contexts are mostly used to verify whether 2 shapes are similar or not. Existing work on shape context requires transformation for shape alignment but this research eliminates this extra work. This paper presents a modified version of shape context for signature verification on bank cheque using K-Nearest Neighbour classifier. Proposed system also demonstrates effective performance when compared with other pattern matching technique Local Binary Pattern.