Block Induced Signature Generative Adversarial Network (BISGAN): Signature Spoofing Using GANs and Their Evaluation

TL;DR

This paper introduces BISGAN, a GAN-based model that generates high-quality forged signatures for biometric spoofing, achieving up to 100% success in fooling signature verification systems and providing a new evaluation method for forgery quality.

Contribution

The work presents a novel generator-focused GAN architecture with attention mechanisms for signature spoofing and introduces a new evaluation technique for forged signature quality.

Findings

Achieves 80-100% success rate in signature spoofing

Uses a CycleGAN-based generator with attention blocks

Provides a new goodness measure for forged signatures

Abstract

Deep learning is actively being used in biometrics to develop efficient identification and verification systems. Handwritten signatures are a common subset of biometric data for authentication purposes. Generative adversarial networks (GANs) learn from original and forged signatures to generate forged signatures. While most GAN techniques create a strong signature verifier, which is the discriminator, there is a need to focus more on the quality of forgeries generated by the generator model. This work focuses on creating a generator that produces forged samples that achieve a benchmark in spoofing signature verification systems. We use CycleGANs infused with Inception model-like blocks with attention heads as the generator and a variation of the SigCNN model as the base Discriminator. We train our model with a new technique that results in 80% to 100% success in signature spoofing. Additionally, we create a custom evaluation technique to act as a goodness measure of the generated forgeries. Our work advocates generator-focused GAN architectures for spoofing data quality that aid in a better understanding of biometric data generation and evaluation.