Novel Approach to Offline Signature Classification and Verification System

Now a days, signatures plays an important role in person identification and verification purpose. There is need of different methods for automatic signature classification as well as for verification system because financial and business related transactions became dependent and authorized via signatures hence in this work, based on a combination of features extracted such as normalized area of signature, aspect ratio, centroid features, trisurface feature, six fold feature and transition feature from the input signatures that is of train as well as from test signature are used for the signature classification and verification purpose. The system is trained using various samples signature of individual and from those various set of signature samples, we extract feature vectors for every indiduals signatures all signature which are used as template sign for the test signature during verification phase. Euclidean distance is used as classifier between the database signature and the test signature.