Deep Learning Methods for Signature Verification

TL;DR

This paper explores deep learning architectures like CNN, RNN, and combined models for signature verification, enhancing feature encoding to improve accuracy in distinguishing genuine signatures from forgeries.

Contribution

It introduces novel deep learning models and improved feature encoding methods specifically designed for signature verification tasks.

Findings

Deep learning models effectively verify signatures.

Enhanced Path Signature Features improve forgery detection.

Models outperform traditional methods in accuracy.

Abstract

Signature is widely used in human daily lives, and serves as a supplementary characteristic for verifying human identity. However, there is rare work of verifying signature. In this paper, we propose a few deep learning architectures to tackle this task, ranging from CNN, RNN to CNN-RNN compact model. We also improve Path Signature Features by encoding temporal information in order to enlarge the discrepancy between genuine and forgery signatures. Our numerical experiments demonstrate the effectiveness of our constructed models and features representations.