Workshop on  Generative, Adversarial and Presentation Attacks In Biometrics (GAPBio) 

Recent advances in generative AI, including Generative Adversarial Networks (GANs), diffusion models, Vision-Language Models (VLMs), and Large Language Models (LLMs), have enabled the creation of highly realistic synthetic images, videos, speech, and multimodal content. While these technologies have transformed computer vision and multimedia applications, they also introduce significant challenges for biometric authentication, digital identity, and multimedia forensics.

Deepfakes, face morphing, presentation attacks, adversarial manipulations, and AI-generated synthetic identities increasingly threaten the reliability of biometric and vision-based recognition systems. At the same time, the rapid development of generative and multimodal models creates new challenges for detecting manipulated content, verifying its authenticity and provenance, and building systems that remain robust against previously unseen attacks.

These challenges are highly relevant to the WACV community, bringing together research at the intersection of computer vision, generative AI, multimodal learning, adversarial machine learning, biometrics, digital forensics, and trustworthy AI. GAPBio provides a dedicated forum to discuss emerging attacks, benchmark new defenses, and identify future research directions for secure and reliable vision and biometric systems. The workshop also aims to foster collaboration between academia, industry, and government.

GAPBio 2027 will be the ninth edition of the Workshop on Generative, Adversarial and Presentation Attacks in Biometrics. Building on eight previous editions held at BTAS 2018 and WACV 2020–2026, the workshop will continue to bring together researchers and practitioners working toward biometric and vision systems that are resilient to next-generation AI-generated attacks.

Papers are invited to report on the following topics, but are not limited to:

• Generative AI attacks on biometric systems (e.g., GANs, diffusion models, VLMs, and LLMs).

• Presentation attack detection (PAD) and biometric spoofing.

• Image, video, and audio manipulation attacks in biometric verification and identification.

• Deepfake, face-swapping, and synthetic identity attacks.

• Face morphing attacks and morph detection.

• Generalizable and foundation model-based attack detection methods.

• Multimodal biometrics and cross-modal authenticity verification.

• VLMs and LLMs for biometrics and digital forensics.

• Multimedia forensics, provenance, and integrity verification of AI-generated content.

• Explainable, trustworthy, and privacy-preserving AI for biometrics.

• Adversarial machine learning and robustness of biometric systems.

• Human-AI collaboration and human perception in authenticity verification.

• Ethical, legal, and societal implications of generative AI in biometrics and digital identity.

• Benchmark datasets, evaluation protocols, and real-world case studies on AI-generated biometric attacks and defenses.

Submission Guidelines

GAPBio 2027 will accept regular archival submissions in line with the WACV 2027 paper guidelines. Accepted regular papers will be included in the official WACV 2027 Workshop Proceedings.

In addition, the workshop will accept non-archival presentation-only submissions, allowing authors to present relevant work while opting out of inclusion in the proceedings.

Detailed submission instructions, paper formatting requirements, templates, and the submission link will be announced.

Camera-ready Submission Guidelines

To be announced.

Important Dates

• Workshop: WACV 2027 — date to be announced

• Full Paper Submission: To be announced

• Acceptance Notification: To be announced

• Camera-Ready Paper: To be announced

Program (to be announced): 

Organizing Committee:

• Kiran Raja, NTNU, Norway

• Naser Damer, Fraunhofer IGD, Germany

• Julian Fiérrez, Universidad Autonoma de Madrid , Spain

• Raghavendra Ramachandra, NTNU, Norway