Can We Leave Deepfake Data Behind in Training Deepfake Detector?
Open accessThe DOI matches its Crossref record; 10 of the 10 references it cites were found in a registry or an index.
Abstract
The generalization ability of deepfake detectors is vital for their applications in real-world scenarios. One effective solution to enhance this ability is to train the models with manually-blended data, which we termed "blendfake", encouraging models to learn generic forgery artifacts like blending boundary. Interestingly, current SoTA methods utilize blendfake without incorporating any deepfake data in their training process. This is likely because previous empirical observations suggest that vanilla hybrid training (VHT), which combines deepfake and blendfake data, results in inferior performance to methods using only blendfake data (so-called "1+1<2"). Therefore, a critical question arises: Can we leave deepfake behind and rely solely on blendfake data to train an effective deepfake detector? Intuitively, as deepfakes also contain additional informative forgery clues (e.g., deep generative artifacts), excluding all deepfake data in training deepfake detectors seems counter-intuitive. In this paper, we rethink the role of blendfake in detecting deepfakes and formulate the process from "real to blendfake to deepfake" to be a progressive transition. Specifically, blendfake and deepfake can be explicitly delineated as the oriented pivot anchors between "real-to-fake" transitions. The accumulation of forgery information should be oriented and progressively increasing during this transition process. To this end, we propose an Oriented Progressive Regularizor (OPR) to establish the constraints that compel the distribution of anchors to be discretely arranged. Furthermore, we introduce feature bridging to facilitate the smooth transition between adjacent anchors. Extensive experiments confirm that our design allows leveraging forgery information from both blendfake and deepfake effectively and comprehensively.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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?
In the citing paper
“This strategy helps the model to learn a more structured representation of forgery artifacts rather than treating each type as disconnected examples.”
From Deepfake detection across image, video, and audio: a comprehensive survey with empirical evaluation of generalization and robustness (2026) · cites this paper as [169]
Citations
Cites
The publisher hasn’t deposited this paper’s reference list.
Cited by
Yan, Z., Zhao, Y., Shen, C., Guo, M., Fu, X., Yao, T., Ding, S., Wu, Y., & Li, Y. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.01177 · Cited by 116
Self-citationX2-DFD: A framework for explainable and extendable Deepfake Detection
Chen, Y., Yan, Z., Cheng, G., Zhao, K., Lyu, S., & Wu, B. (2025) · neural information processing systems · doi:10.52202/085713-2809 · Cited by 57
Self-citationDeepfake Detection that Generalizes Across Benchmarks
Yermakov, A., Čech, J., Matas, J., & Fritz, M. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00082 · Cited by 68
Nguyen-Le, H.-H., Tran, V.-T., Nguyen, D.-T., & Le-Khac, N.-A. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-026-11608-4 · Cited by 28
Kundu, R., Xiong, H., Mohanty, V., Balachandran, A., & Roy–Chowdhury, A. K. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02612 · Cited by 57
Metadata
- Authors
- Cheng, J., Yan, Z., Zhang, Y., Luo, Y., Wang, Z., & Li, C.
- Venue
- Advances in Neural Information Processing Systems 37
- Published
- 2024
- Pages
- 21979-21998
- ISSN
- 1049-5258 (ISSN-L)
- OpenAlex
- W4415798721
- Cited by
- 29
- References
- 0
- Citations by year
- 2026: 25 · 2025: 4
- FWCI
- 6.01
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- public-domain
- Locations
- neural information processing systems (publishedVersion)
- arXiv (Cornell University) (submittedVersion, public-domain)
- arXiv (Cornell University) (public-domain)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
- Physical Sciences › Computer Science › Signal Processing › Biometric Identification and Security
- Keywords
- deepfake detection, synthetic data, data bias, generative adversarial networks, training data quality
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
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.