Rapid off-line signature verification based on Signature Envelope and Adaptive Density Partitioning

Abstract

Handwritten signature is a widely used biometric which incorporates high intra personal variance. The most challenging problem in automatic signature verification is to extract features which are robust against this natural variability and at the same time discriminate between genuine and fake samples. This paper presents a novel method for extracting easily computed rotation and scale invariant features for offline signature verification. These features are extracted using the signature envelope and adaptive density partitioning. The effectiveness of the proposed features has been investigated over 900 signatures using a neural network classifier. The experimental results show the verification accuracy rate of 90.7%.