CoChat

DeepFake-Adapter: Dual-Level Adapter for DeepFake Detection

Open access
Shao, Rui, Wu, Tianxing, Nie, Liqiang, Liu, Ziwei · 2025 · International Journal of Computer Vision

The DOI matches its Crossref record; 84 of the 84 references it cites were found in a registry or an index.

Abstract

Existing deepfake detection methods fail to generalize well to unseen or degraded samples, which can be attributed to the over-fitting of low-level forgery patterns. Here we argue that high-level semantics are also indispensable recipes for generalizable forgery detection. Recently, large pre-trained Vision Transformers (ViTs) have shown promising generalization capability. In this paper, we propose the first parameter-efficient tuning approach for deepfake detection, namely DeepFake-Adapter, to effectively and efficiently adapt the generalizable high-level semantics from large pre-trained ViTs to aid deepfake detection. Given large pre-trained models but limited deepfake data, DeepFake-Adapter introduces lightweight yet dedicated dual-level adapter modules to a ViT while keeping the model backbone frozen. Specifically, to guide the adaptation process to be aware of both global and local forgery cues of deepfake data, 1) we not only insert Globally-aware Bottleneck Adapters in parallel to MLP layers of ViT, 2) but also actively cross-attend Locally-aware Spatial Adapters with features from ViT. Unlike existing deepfake detection methods merely focusing on low-level forgery patterns, the forgery detection process of our model can be regularized by generalizable high-level semantics from a pre-trained ViT and adapted by global and local low-level forgeries of deepfake data. Extensive experiments on several standard deepfake detection benchmarks validate the effectiveness of our approach. Notably, DeepFake-Adapter demonstrates a convincing advantage under cross-dataset and cross-manipulation settings.

Overview

Integrity signals

Registry

Crossref · 10.1007/s11263-024-02274-6

Cites

84 of 84 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Verification

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 84 references

Research works printed but not linked: 13

  1. Agarap, A.F. (2018). Deep learning using rectified linear units (relu). arXiv preprint arXiv:1803.08375. · arXiv
  2. Ba, J.L., Kiros, J.R., & Hinton, G.E. (2016). Layer normalization. arXiv preprint arXiv:1607.06450. · arXiv
  3. Chen, Z., Duan, Y., Wang, W., He, J., Lu, T., Dai, J., & Qiao, Y. (2023). Vision transformer adapter for dense predictions. In: ICLR.
  4. Dhariwal, P., & Nichol, A. (2021). Diffusion models beat gans on image synthesis. NeurIPS, 34, 8780–8794.
  5. Dolhansky, B., Bitton, J., Pflaum, B., Lu, J., Howes, R., Wang, M., & Ferrer, C.C. (2020). The deepfake detection challenge (dfdc) dataset. arXiv preprint arXiv:2006.07397. · arXiv
  6. Ho, J., Jain, A., & Abbeel, P. (2020). Denoising diffusion probabilistic models. In: NeurIPS, 33, 6840–6851.
  7. Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., & Chen, W. (2022). Lora: Low-rank adaptation of large language models. In: ICLR.
  8. Liu, L., Ren, Y., Lin, Z., & Zhao, Z. (2022). Pseudo numerical methods for diffusion models on manifolds. In: ICLR.
  9. Van der Maaten, L., & Hinton, G. (2008). Visualizing data using t-sne. Journal of machine learning research 9(11).
  10. Quinn, N.A., & Dhariwal, P. (2021). Improved denoising diffusion probabilistic models. In: ICML, pp. 8162–8171.
  11. Shao, R., Wu, T., & Liu, Z. (2023). Robust sequential deepfake detection. arXiv preprint arXiv:2309.14991. · arXiv
  12. Tran, D., Bourdev, L.D., Fergus, R., Torresani, L., & Paluri, M. (2014). C3d: generic features for video analysis. CoRR, abs/1412.0767. · arXiv
  13. Wang, G., Zhou, J., & Wu, Y. (2020). Exposing deep-faked videos by anomalous co-motion pattern detection. arXiv preprint arXiv:2008.04848. · arXiv

Cited by

See all 49

Metadata

Authors
Shao, R., Wu, T., Nie, L., & Liu, Z.
Affiliations
  • Harbin Institute of Technology (CN)
  • Nanyang Technological University (SG)
Corresponding author
Liu, Ziwei
Venue
International Journal of Computer Vision
Published
2025-01-24
Volume
133
Issue
6
Pages
3613-3628
ISSN
  • 0920-5691 (ISSN-L)
  • 1573-1405
OpenAlex
W4406804129
Cited by
48
References
79
Citations by year
2026: 21 · 2025: 22
FWCI
15.10
Type
journal article
Language
en
Open access
Licence
Locations
  • International Journal of Computer Vision (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • DR-NTU (Nanyang Technological University) (submittedVersion)
  • arXiv (Cornell University)
Topics
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Generative Adversarial Networks and Image Synthesis
  • Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
Keywords
deepfake detection, parameter-efficient fine-tuning, vision transformer, high-level semantics, generalization to unseen data, cross-dataset generalization, frozen backbone

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: DeepFake-Adapter: Dual-Level Adapter for DeepFake Detection. https://app.cochat.ai/papers/10.1007/s11263-024-02274-6
The paper
Shao, R., Wu, T., Nie, L., & Liu, Z. (2025). DeepFake-Adapter: Dual-Level Adapter for DeepFake Detection. International Journal of Computer Vision. https://doi.org/10.1007/s11263-024-02274-6

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.