Vital Sign: Personal Signature Based Biometric Authentication System

Signature is a behavioral biometric: it is not based on the physical properties, such as fingerprint or face, of the individual, but behavioral ones. Signature verification is split into two according to the available data in the input. Offline (static) signature verification takes as input the image of a signature and is useful in automatic verification of signatures found on bank checks and documents. Online (dynamic) signature verification uses signatures that are captured by pressure-sensitive tablets that extract dynamic properties of a signature in addition to its shape. The purpose of project is to develop an authentication system based on personal signatures. Signature verification is an important research topic in the area of biometric verification. In this project the work is done in such a way that the signatures are captured using pen tablet. The signatures are gathered using local features and 1-D Wavelet transform (WT) is performed on that features. The Discrete Cosine Transform (DCT) is used to reduce the approximation coefficients vector obtained by WT to a feature vector of a given dimension. Then the classification is performed using Linear programming descriptor (LPD), Simple K-Nearest Neighbor (kNN) and Pruned fuzzy k-Nearest Neighbor (Pwfknn). The results obtain in this projects are with improved knn FAR and FRR for random forgeries is 14.40% and 3.2% respectively and FAR for skilled forgeries 12.80%, with lpd FAR and FRR on random forgeries is 1.52% and 23.80% respectively and FAR for skilled forgeries is 14.20%, with pwfknn FAR 3.44% and FRR 13.11% on random and FAR of 15.4%, on skilled forgeries.