发表机构
Graduate School of Information Sciences, Tohoku University（东北大学信息科学研究生院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对人脸反欺骗需求，提出DINO-VPT轻量级仅视觉框架，利用分层视觉提示调整，通过提示路由网络动态注入提示，无需多模态融合，在基准测试中比现有方法准确率更高，证明合理架构能达最优性能。
AI 中文摘要
随着欺骗攻击的日益多样化，对能够检测物理和数字威胁的统一人脸反欺骗（FAS）模型的需求不断增长。虽然现有的视觉语言模型（VLM）在此背景下具有较高的泛化能力，但它们严重依赖复杂的多模态融合和外部文本编码器。本文提出了DINO-VPT，一个轻量级的、仅视觉的框架，利用分层视觉提示调整。通过提示路由网络（PRN）根据输入特征动态注入提示，该方法无需多模态融合就能有效解开各种欺骗伪像。在UniAttackData基准上的评估表明，DINO-VPT比基于VLM的现有方法具有更高的准确率。结果表明，结构合理的仅视觉架构无需多模态监督就能在统一FAS中达到最优性能。
英文摘要
With the increasing diversity of spoofing attacks, there is a growing demand for unified Face Anti-Spoofing (FAS) models capable of detecting both physical and digital threats. While existing Vision-Language Models (VLMs) demonstrate high generalization in this context, they heavily rely on complex multimodal fusion and external text encoders. In this paper, we propose DINO-VPT, a lightweight, vision-only framework leveraging hierarchical visual prompt tuning. By dynamically injecting prompts conditioned on input features via a Prompt Routing Network (PRN), our method effectively disentangles diverse spoofing artifacts without requiring multimodal fusion. Evaluations on the UniAttackData benchmark demonstrate that DINO-VPT achieves higher accuracy than state-of-the-art VLM-based methods. Our results indicate that a properly structured vision-only architecture can achieve state-of-the-art performance in unified FAS without the need for multimodal supervision.
Commentsaccepted to IJCB2026
