IMAGE PROCESSING BASED ON OFFLINE HANDWRITTEN SIGNATURE RECOGNITION AND VERIFICATION SCHEME

Abstract

In today's world, numerous techniques are used for human identification. These can be classified into vision-based and non-vision based. Vision-based are an eye, face and fingerprint recognition and iris and retina scanning, whereas the most widely used non-vision based techniques are voice recognition and signature recognition and verification. Signature has remained pivotal bio-metric trait as it can be used for a person's identity as every person has a different signature from that of the other, which distinctive characteristic is either physiological or behavioral.  Handwritten signature serves as ''seal of approval'' and is widely used mode for authentication purpose in fields like finance, education, legal transaction, and banking etc. Online i.e dynamic and offline i.e static are the two key types of signature recognition and verification techniques.  Image processing technique has been used in this study for offline signature identification. Under this system image of a person's signature is obtained and its recognition is made through several techniques including ANN. For training of the system signature as an image is taken as a sample and checked with already stored samples, which requires authentication. Many kinds of research have been carried out on different aspects of this system yielding distinctive results, which aspects are image feature extraction; modified NN approach; global feature extraction using ANN and training. In this paper, a brief description of offline handwritten signature recognition and verification scheme are disclosed.