Offline Signature Recognition Using Delaunay Triangulation

Abstract

Biometric identification is evolving technology solving problem related to access control and identity verification. Signature can be used in various places to keep out thieves and intruder. For secure personal verification system offline signature has been recognized as a key competitor. However its replication has become a serious risk, fabricating various serious risks to access control and personal identification. To mitigate this risk, we propose a novel method to detect skilled forgery, consisting of some preprocessing steps, feature extraction, Delaunay triangulation, normalization and recognition. The aim of the proposed method is to decrease false acceptance rate and false rejection rate which is often used in signature verification method. The precision of recognizing genuine signature of the proposed method is 96% but rotation of the sample signature image decrease the desired result slightly. Comparative analysis of proposed method with other state-of-the-art techniques has been made in the experimental section, verifying its rationality.