DINO-VPT: Hierarchical Visual Prompt Tuning for Joint Physical-Digital Face Anti-Spoofing
Quick Answer
DINO-VPT introduces a lightweight, vision-only framework for face anti-spoofing, achieving superior accuracy over existing VLM-based methods on the UniAttackData benchmark without multimodal fusion.
Quick Take
This approach utilizes hierarchical visual prompt tuning to effectively address diverse spoofing threats, demonstrating that vision-only architectures can excel in unified FAS tasks.
Key Points
- DINO-VPT leverages hierarchical visual prompt tuning for face anti-spoofing.
- Achieves higher accuracy than state-of-the-art methods on UniAttackData.
- No multimodal fusion or external text encoders are required.
- Demonstrates effectiveness against both physical and digital spoofing attacks.
- Accepted for presentation at IJCB2026.
Paper Resources
📖 Reader Mode
~2 min readAbstract: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.
| Comments: | accepted to IJCB2026 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2607.20900 [cs.CV] |
| (or arXiv:2607.20900v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2607.20900 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Koichi Ito Dr [view email]
[v1]
Thu, 23 Jul 2026 03:51:22 UTC (5,793 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.