DGPDL: Domain-Guided Prompt Distribution Learning for Generalizable Face Anti-Spoofing
IEEE Transactions on Pattern Analysis and Machine Intelligence · 2026-08 · 48 (8) · 8953-8969
DOI: 10.1109/tpami.2026.3674204 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2026-08. Verfasst von Ajian Liu und 8 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 93 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Beijing Natural Science Foundation, Science and Technology Development Fund of Macau.
Abstract (Originalsprache)
The overfitting of domain signals results in poor domain generalization of face anti-spoofing. The current methods usually improve the diversity of source domains to alleviate this overfitting. However, this benefit is minimal, as even the most diverse domain signals will also be absent in the target domain. In this work, we propose a Domain-Guided Prompt Distribution Learning (DGPDL) built on Vision-Language Models like CLIP, which explores a unified representation of domain signals as a prompt across the source and target domain to alleviate the understanding bias caused by domain gaps. Specifically, we first define a learnable Domain-Specific Distribution (DSD) that covers as many domain elements as possible, such as image quality, color tone, camera settings, etc., which establish connections between different domains and linearly combinable prompt in any domain; Then, based on the style statistics of the given sample, we construct its optimal Domain-Specific Prompts (DSPs) from the defined DSD through the designed Prompt Assemble Attention (PAA) with the similarity matching; Finally, the assembled DSPs will act as carrier or agent to perform on both the vision and language branches, synergistically improving the model's recognition of domain signals. By using the prompt to represent domain signals uniformly, if the model can be robust to DSPs in the source domain, it should be applicable to target domain, as they share the same DSD. By representing domain signals as prompts rather than instantiation features, DGPDL effectively reduces the reliance on specific domain appearances. This design enables the model to dynamically adapt to unseen target domains without the need for retraining. Extensive experiments show that the DGPDL is effective and outperforms the state-of-the-art methods on several cross-domain benchmarks.
Förderung
- National Natural Science Foundation of China
- Beijing Natural Science Foundation
- Science and Technology Development Fund of Macau
- Medical Science and Technology Foundation of Guangdong Province
- Open Research Fund of The State Key Laboratory of Blockchain and Data Security, Zhejiang University
- AI Super Computing Platform of Macau University of Science and Technology
Zitieren
APA
Liu, A., Lin, X., Zhi, R., Liang, Y., Zhu, X., Cai, Z., Wan, J., Escalera, S., & Lei, Z. (2026). DGPDL: Domain-Guided Prompt Distribution Learning for Generalizable Face Anti-Spoofing. IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(8), 8953-8969. https://doi.org/10.1109/tpami.2026.3674204
Vancouver
Liu A, Lin X, Zhi R, Liang Y, Zhu X, Cai Z, et al. DGPDL: Domain-Guided Prompt Distribution Learning for Generalizable Face Anti-Spoofing. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2026;48(8):8953-8969. doi:10.1109/tpami.2026.3674204
BibTeX
@article{Liu2026,
author = {Liu, Ajian and Lin, Xun and Zhi, Ruicong and Liang, Yanyan and Zhu, Xinshan and Cai, Zhanchuan and Wan, Jun and Escalera, Sergio and Lei, Zhen},
title = {DGPDL: Domain-Guided Prompt Distribution Learning for Generalizable Face Anti-Spoofing},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
year = {2026},
volume = {48},
number = {8},
pages = {8953--8969},
doi = {10.1109/tpami.2026.3674204}
}
EUVIMED