RCFA 2026

Workshop and Challenge on Robust Continuous Face Authentication

(RCFA 2026)

Associated with the Asian Conference on Computer Vision (ACCV 2026), Osaka, Japan, 15 Dec. 2026

Introduction

Continuous face authentication aims to continuously verify a user's identity throughout an interaction, rather than performing a one-time authentication from a single image or a short video clip. This capability is becoming increasingly important for many real-world applications, including online meetings, online banking, remote examinations, telemedicine, and secure human-computer interaction.

Despite the remarkable progress of face recognition, continuous face authentication has received much less attention than conventional image-based verification. Compared with static face recognition, continuous authentication must cope with long-term interactions and presents additional challenges, including large intra-subject variations (e.g., motion blur, illumination changes, low-quality frames, and occlusions such as face masks) as well as security threats (e.g., identity transition attacks and AI-generated face attacks).

The RCFA 2026 Challenge aims to advance research on robust and secure continuous face authentication by providing a standardized benchmark for realistic evaluation. We hope this benchmark will promote future research in continuous biometric authentication.

Challenge Description

The challenge consists of two tracks: Robustness and Security.

(1) The Robustness track evaluates continuous authentication under long-term appearance variations, including motion blur, illumination changes, low-quality frames, and facial occlusions such as masks.

(2) The Security track focuses on detecting identity-related attacks, including identity-transition attacks caused by facial tracker hijacking and digital facial attacks generated using modern generative AI techniques, such as diffusion-based face synthesis and identity-preserving face generation models.

Codabench website: [Link]

# Winning teams will receive official award certificates and will be invited to present their approaches and results at the workshop, either on site or online. This offers participants an opportunity to showcase their work, exchange ideas, and connect with researchers in biometrics, computer vision, and security.