Robust Offline Signature Verification Based On Global Features

Abstract

Biometrics is a verification system that identifies a person based on his physiological and behavioral features. Signature verification is the biometrics identification method which is legally accepted and used in many commercial fields such as e-business, access control and so on. In this paper we propose a robust off-line signature verification based on global features (ROSVGF) for skilled and random forgeries. In this model prior to extracting the features, we preprocessed the signatures in the database. Preprocessing consists of i) normalization ii) noise reduction iii) thinning and skelitazition, for feature set extraction which consists of global features such as signature height-to-width ratio (aspect ratio), maximum horizontal histogram and maximum vertical histogram, horizontal center and vertical center of the signature, end points of the signature, signature area. It is observed that our proposed model gives the better Type I and Type II errors compared to existing models.