Introduction

In recent years, the growing sophistication of face spoofing techniques has significantly threatened the security of face recognition systems. Face Anti-Spoofing (FAS) has thus become a critical component in ensuring the reliability of biometric authentication. To stimulate research in this area, we have organized five successful editions of the Face Anti-Spoofing Workshop and Challenge at major conferences, including CVPR 2019, CVPR 2020, ICCV 2021, CVPR 2023, and CVPR 2024, attracting over 1,400 participating teams from academia and industry. However, unified detection of both physical and digital attacks remains a major challenge. While physical presentation attacks such as print, replay, and 3D mask attacks typically introduce artifacts like color distortion and moiré patterns, digital forgeries—including identity manipulations, adversarial examples, and GAN-based generation—alter facial imagery at the pixel level in often imperceptible ways. Existing solutions still treat these as separate tasks, hindering the development of generalizable models. To address this gap, the 6th Face Anti-Spoofing Workshop@ICCV 2025 introduces the task of Unified Physical-Digital Attack Detection and releases a significantly expanded dataset, UniAttackData+. We collected data from 2,875 participants representing three distinct ethnic groups (i.e., African, East Asian, and Central Asian), capturing 18,250 authentic videos under various lighting conditions, backgrounds, and acquisition devices. For each participant, we applied 54 different attack methods, including 14 physical attacks and 40 digital attacks, resulting in a total of 679,097 forged videos.

Challenge website: [Link] 

Workshop Paper Submission Link:  Waiting for opening......

Ref:

[1] Ajian Liu, Haocheng Yuan, Xiao Guo, Hui Ma, Wanyi Zhuang, Changtao Miao, Jun Lan, Qi Chu, Jun Wan, Xiaoming Liu,  Zhen Lei Benchmarking Unified Face Attack Detection via Hierarchical Prompt Tuning.  [Link]

Workshop Schedule 

Date: Oct. 19th, 2025

Time: 8:30 AM - 12:00 AM 

Location: Arch 201

• 8:30 - 8:40 Opening of contest, overview of results.

• 8:40 - 9:15 Keynote Speaker 1.

• 9:15 - 9:30 Challenge champion presentation.

• 9:30 - 9:45 Challenge second place presentation.

• 9:45 - 10:20 Keynote Speaker 2.

• 10:20 - 10:35 Challenge third place presentation.

• 10:35 - 10.40 Coffee break.

• 10.40 - 11.15 Keynote Speaker 3.

• 11:15 - 11:30 Challenge invited team presentations.

• 11:30 - 11:45 Awards Ceremony.

• 11:45 - 12:00 Closing.

Important Competition Dates:

• 20 May 2025:​​ Quantitative competition begins; Development data released: training set with true labels; Validation data released: validation set without true labels; Encrypted test data released (no labels provided).

• ​​12 June 2025:​​ Release of encrypted final evaluation data (test data without true labels); True labels of validation set released.

• ​​13 June 2025:​​ Release of final evaluation data decryption key; Participants start predicting results on the final evaluation data.

• ​​28 June 2025:​​ Quantitative competition ends; Deadline for submitting final predictions on test data; Deadline for code submission (organizers verify code on test data).

• ​​30 June 2025:​​ Challenge results officially released.

• ​​9 July 2025:​​ Deadline for workshop paper submissions.

• ​​11 July 2025:​​ Deadline for authors to submit meta-data of accepted papers.

• ​​18 August 2025:​​ Camera-ready paper submission deadline.

• ​​19 October 2025:​​ Workshop@ICCV 2025 held; Challenge results announced; Award ceremony conducted.