OPTIMIZED MULTI MODEL SYSTEM FOR ONLINE SIGNATURE VERIFICATION

Abstract

Handwritten signature is the most widely accepted biometric to identity verification. The proposed online handwritten signature verification system consists mainly of three phases: Signal preprocessing, feature extraction, and feature matching. Steps for verifying online handwritten signature in this system start with extracting dynamic data (x and y positions) of points that forming the signature. Pen-movement angles and speed are then derived from pen position data. To reduce variations in pen-position and pen-movement angles dimensionality, data is normalized. After that all the parameters are to be put in a single vector. Here in the proposed system five parameters are taken in to account. Features of the signature can be extracted using proposed feature extraction method. Corresponding to every signature a unique feature will be extracted and this will be quantized using quantization step size vector. Both the feature vector and quantization vector are to be stored using template generator. Further, during the matching phase, a different distance measuring algorithm has been implemented known as Minkowski distance which is helpful in improving the results.