Verification of persons via face and signature analysis

Abstract

This article proposes a new approach for verification of people. The model consists of two parts: face and signature analysis. For face information processing morphological filtering is used to enhance the intrinsic features of a face, reduce the influence of rotation in depth, changes in facial expression, hair style, glasses and lighting conditions. The filtered images are then a subject for learning by a modified high order neural network. In signature analysis the model first traces the signature to extract the dynamical information that is usually lost in an off-line mode. After this step a neural network (neocognitron with switching attention) is used to recognize and finally verify the signature. These two parts can work independently and finally their outputs can be used to form a complex person verifier.