English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

While previous studies have explored attacks via random, simple, and skilledforgeries, generative attacks have received limited attention in thedata-driven signature verification (DASV) process. Thus, this paper exploresthe impact of generative attacks on DASV and proposes practical andinterpretable countermeasures. We investigate the power of two prominent DeepGenerative Models (DGMs), Variational Auto-encoders (VAE) and ConditionalGenerative Adversarial Networks (CGAN), on their ability to generate signaturesthat would successfully deceive DASV. Additionally, we evaluate the quality ofgenerated images using the Structural Similarity Index measure (SSIM) and usethe same to explain the attack's success. Finally, we propose countermeasuresthat effectively reduce the impact of deep generative attacks on DASV.  We first generated six synthetic datasets from three benchmarkoffline-signature datasets viz. CEDAR, BHSig260- Bengali, and BHSig260-Hindiusing VAE and CGAN. Then, we built baseline DASVs using Xception, ResNet152V2,and DenseNet201. These DASVs achieved average (over the three datasets) FalseAccept Rates (FARs) of 2.55%, 3.17%, and 1.06%, respectively. Then, we attackedthese baselines using the synthetic datasets. The VAE-generated signaturesincreased average FARs to 10.4%, 10.1%, and 7.5%, while CGAN-generatedsignatures to 32.5%, 30%, and 26.1%. The variation in the effectiveness ofattack for VAE and CGAN was investigated further and explained by a strong (rho= -0.86) negative correlation between FARs and SSIMs. We created another set ofsynthetic datasets and used the same to retrain the DASVs. The retainedbaseline showed significant robustness to random, skilled, and generativeattacks as the FARs shrank to less than 1% on average. The findings underscorethe importance of studying generative attacks and potential countermeasures forDASV.

Deep Generative Attacks and Countermeasures for Data-Driven Offline Signature Verification