pith. sign in

arxiv: 2512.04331 · v2 · pith:BRHABJKAnew · submitted 2025-12-03 · 💻 cs.CV

Open Set Face Forgery Detection via Dual-Level Evidence Collection

Zhongyi Cai , Bryce Gernon , Wentao Bao , Yifan Li , Matthew Wright , Yu Kong This is my paper

Pith reviewed 2026-05-21 17:41 UTC · model grok-4.3

classification 💻 cs.CV
keywords open set face forgery detectionuncertainty estimationdeepfake detectionspatial frequency analysisevidential learning
0
0 comments X

The pith

Dual-level evidence collection allows detectors to identify novel face forgery categories by estimating uncertainty at spatial and frequency levels.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper studies open set face forgery detection, where a model must spot entirely new fake categories that did not appear in training data. It reformulates the task as uncertainty estimation and introduces the DLED method that gathers category-specific evidence separately from spatial image details and from frequency components before combining them. A sympathetic reader would care because face generation tools keep changing, so detectors limited to known fakes quickly become useless in practice. Experiments show the method outperforms prior approaches especially when facing these unseen forgery types. It also stays competitive on ordinary real-versus-fake checks.

Core claim

The paper claims that reformulating open set face forgery detection as an uncertainty estimation problem and solving it with dual-level evidence collection at spatial and frequency domains enables reliable identification of novel fake categories, as shown by state-of-the-art results that exceed baseline models by a 20 percent margin on average for unseen fakes while remaining competitive on standard binary detection.

What carries the argument

Dual-Level Evidential face forgery Detection (DLED) that extracts and integrates category-specific evidence on the spatial and frequency levels to estimate prediction uncertainty.

If this is right

  • Detectors using this approach can flag forgeries from forgery categories never seen during training.
  • Performance gains reach an average 20 percent margin over existing baselines specifically on novel fake categories.
  • The method still delivers competitive accuracy on the standard binary real-versus-fake classification task.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same dual-level evidence idea could be tested on video or audio forgeries to see whether uncertainty signatures remain useful outside still images.
  • If frequency-domain evidence consistently drives the uncertainty signal, detectors might be simplified by focusing resources there for new forgery types.
  • Real-world deployment would benefit from checking how the method behaves when a single image contains a mix of known and unknown forgery artifacts.

Load-bearing premise

The reformulation of the OSFFD problem through uncertainty estimation assumes that novel forgery categories will reliably produce distinguishable uncertainty signatures when evidence is collected separately at spatial and frequency levels.

What would settle it

A new forgery generation algorithm whose output images receive low uncertainty scores indistinguishable from those of known real or fake categories would show the dual-level signatures do not reliably flag novel fakes.

Figures

Figures reproduced from arXiv: 2512.04331 by Bryce Gernon, Matthew Wright, Wentao Bao, Yifan Li, Yu Kong, Zhongyi Cai.

Figure 1
Figure 1. Figure 1: Comparison with existing settings. Different from DeepFake Detection (a) and Attribution (b), Open Set Face Forgery Detection (c) aims to identify whether a forgery originates from a novel fake category or not while simultaneously performing multiclass classification among real and known fake categories. ously identifies emerging, unknown fake categories and per￾forms multiclass classification among real a… view at source ↗
Figure 2
Figure 2. Figure 2: Illustration for Fake Categories in OSFFD. Real faces and fake faces from the seen categories are used to train the model. Subsequently, the model is evaluated on test data that includes both seen classes and previously unseen categories. In the figure, the labels EFS, FR, and FS denote seen categories, whereas FE repre￾sents an unseen category. 2.3. Open Set Recognition Open Set Recognition is a well-defi… view at source ↗
Figure 3
Figure 3. Figure 3: Overview of DLED. DLED collects and fuses evidence from both the spatial and frequency domains to estimate prediction uncertainty. Our improved uncertainty estimation uˆ is applied to achieve better detection performance. FN represents the N-th fake category and K is the total known class number. If the uncertainty for the given sample is larger than the computed threshold, its label will be reassigned to … view at source ↗
Figure 4
Figure 4. Figure 4: Visualization of Attention Map. Attention maps gen￾erated by the DLED model for the novel fake categories FS and FE are shown in subfigures (a) and (b), respectively. To provide a clearer understanding of DLED’s behavior in OSFFD, we present visualizations of attention map in [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Visualization of Evidence Distribution. The evidence for seen fake categories FR and EFS is condensed in their corre￾sponding corner with low uncertainty, while the evidence for novel fake categories FS and FE is sparse with higher uncertainty. ognize newly emerging fake categories while simultane￾ously maintaining strong performance on known classes. Furthermore, we illustrate how DLED distinguishes be￾tw… view at source ↗
Figure 6
Figure 6. Figure 6: Visualization of Evidence Distribution for novel real and fake face images. We visualize the Evidence Distribution of novel real and fake faces and present the corresponding predic￾tions made by our DLED model. The prediction confidence for new real faces remains high, whereas novel fake faces exhibit low confidence accompanied by high prediction uncertainty. 6.5. Influence of Uncertainty Threshold Uncerta… view at source ↗
read the original abstract

The surge in face forgeries has increasingly undermined confidence in the authenticity of online content. As generation algorithms rapidly evolve, new fake categories will constantly emerge, severely challenging existing face forgery detection methods. Although face forgery detection has recently improved, current techniques remain largely confined to binary Real-vs-Fake classification or the recognition of known fake categories. Moreover, they fail to identify the emergence of entirely new forgery methods. In this work, we study the Open Set Face Forgery Detection (OSFFD) problem, which requires the detection model to identify novel fake categories. To enhance its real-world applicability, we reformulate the OSFFD problem and address it through uncertainty estimation. Specifically, we propose the Dual-Level Evidential face forgery Detection (DLED) approach, which estimates prediction uncertainty by extracting and integrating category-specific evidence on the spatial and frequency levels. Comprehensive experiments across diverse settings demonstrate that our proposed DLED approach achieves state-of-the-art performance. Notably, it surpasses various existing baseline models by a $20\%$ margin on average when identifying forgeries from novel fake categories. Concurrently, our DLED method yields competitive performance on the standard binary Real-versus-Fake face forgery detection task.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript studies the Open Set Face Forgery Detection (OSFFD) problem, reformulating it as an uncertainty estimation task. It proposes Dual-Level Evidential face forgery Detection (DLED), which collects and integrates category-specific evidence at spatial and frequency levels to identify novel forgery categories while maintaining competitive binary Real-vs-Fake performance. The central empirical claim is state-of-the-art results with an average 20% improvement over baselines on novel fake categories.

Significance. If the reported gains prove reproducible and the dual-level uncertainty mechanism reliably separates novel forgeries, the work would provide a practical advance for real-world forgery detectors that must handle rapidly evolving generation methods. The explicit reformulation via uncertainty estimation and the spatial-frequency evidence integration constitute a concrete, testable contribution that could be extended to other open-set vision tasks.

major comments (2)
  1. [Abstract] Abstract: the claim of a 20% average improvement on novel categories is stated without any reference to the datasets used, the exact baseline implementations, the number of trials, error bars, or statistical tests. This information is load-bearing for the central performance claim and its absence prevents verification of the reported margin.
  2. [Reformulation and DLED description] Reformulation paragraph and DLED description: the approach assumes that frequency-level evidence (FFT/DCT-based) will produce systematically higher uncertainty for unseen forgery methods than for known ones. No ablation, feature visualization, or analysis is provided to confirm that the frequency artifacts are forgery-category-specific rather than generic image statistics; if the latter holds, the integrated uncertainty score would not reliably flag true open-set cases and the 20% gain would not follow.
minor comments (1)
  1. [Abstract] The abstract refers to 'comprehensive experiments across diverse settings' without even a high-level summary of those settings, which reduces clarity for readers.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments on our manuscript. We have addressed each major point below, indicating where revisions have been made to strengthen the presentation and supporting analysis.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim of a 20% average improvement on novel categories is stated without any reference to the datasets used, the exact baseline implementations, the number of trials, error bars, or statistical tests. This information is load-bearing for the central performance claim and its absence prevents verification of the reported margin.

    Authors: We agree that the abstract would benefit from additional context to support verification of the central empirical claim. In the revised manuscript we have updated the abstract to name the primary datasets (FaceForensics++, Celeb-DF and the open-set splits described in Section 4), the main baseline families, and to note that the reported 20 % average margin is computed across multiple random seeds with standard deviations and full tables provided in the experimental section. Space constraints preclude exhaustive statistical tests in the abstract itself, but the main text now explicitly references the relevant tables and significance checks. revision: yes

  2. Referee: [Reformulation and DLED description] Reformulation paragraph and DLED description: the approach assumes that frequency-level evidence (FFT/DCT-based) will produce systematically higher uncertainty for unseen forgery methods than for known ones. No ablation, feature visualization, or analysis is provided to confirm that the frequency artifacts are forgery-category-specific rather than generic image statistics; if the latter holds, the integrated uncertainty score would not reliably flag true open-set cases and the 20% gain would not follow.

    Authors: We thank the referee for highlighting this important point. While the overall performance gains on novel categories provide indirect support for the utility of the frequency branch, we acknowledge that a direct demonstration of category-specific uncertainty behavior was missing. In the revised manuscript we have added a new subsection (4.3) containing (i) an ablation isolating the frequency-level evidence, (ii) uncertainty histograms comparing known versus novel forgery classes, and (iii) frequency-domain feature visualizations that illustrate distinct artifact patterns for unseen methods. These additions confirm that the frequency evidence contributes forgery-category-specific uncertainty rather than merely reflecting generic image statistics. revision: yes

Circularity Check

0 steps flagged

No circularity: DLED is a novel uncertainty-based reformulation with independent content

full rationale

The paper introduces OSFFD and addresses it by reformulating as uncertainty estimation, then proposes DLED to extract and integrate category-specific evidence at spatial and frequency levels. No equations, fitted parameters renamed as predictions, or self-citation chains are visible that would make the reported 20% margin on novel categories reduce by construction to the inputs or prior results. The method is presented as a new dual-level evidence collection procedure rather than a re-expression of existing quantities, and the central performance claim rests on experimental comparison to external baselines rather than internal redefinition. This qualifies as a self-contained proposal.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Review is limited to the abstract; no explicit free parameters, axioms, or invented entities are named. The core idea rests on the unstated domain assumption that uncertainty derived from dual-level evidence will separate novel forgeries from known classes.

axioms (1)
  • domain assumption Uncertainty estimation via integrated spatial and frequency evidence can identify novel forgery categories
    Central to the reformulation of OSFFD and the DLED method described in the abstract.

pith-pipeline@v0.9.0 · 5748 in / 1324 out tokens · 64917 ms · 2026-05-21T17:41:11.109853+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

77 extracted references · 77 canonical work pages · 3 internal anchors

  1. [1]

    Evidential deep learn- ing for open set action recognition

    Wentao Bao, Qi Yu, and Yu Kong. Evidential deep learn- ing for open set action recognition. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 13349–13358, 2021. 3, 4, 5

    work page 2021

  2. [2]

    End-to-end reconstruction- classification learning for face forgery detection

    Junyi Cao, Chao Ma, Taiping Yao, Shen Chen, Shouhong Ding, and Xiaokang Yang. End-to-end reconstruction- classification learning for face forgery detection. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4113–4122, 2022. 1, 2

    work page 2022

  3. [3]

    Adversarial reciprocal points learning for open set recognition.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(11):8065–8081, 2021

    Guangyao Chen, Peixi Peng, Xiangqian Wang, and Yonghong Tian. Adversarial reciprocal points learning for open set recognition.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(11):8065–8081, 2021. 3

    work page 2021

  4. [4]

    Open-set deep- fake detection to fight the unknown

    Michael Macedo Diniz and Anderson Rocha. Open-set deep- fake detection to fight the unknown. InICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 13091–13095. IEEE,

    work page 2024

  5. [5]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

    Alexey Dosovitskiy. An image is worth 16x16 words: Transformers for image recognition at scale.arXiv preprint arXiv:2010.11929, 2020. 7

    work page internal anchor Pith review Pith/arXiv arXiv 2010

  6. [6]

    Flexible visual recognition by evidential modeling of con- fusion and ignorance

    Lei Fan, Bo Liu, Haoxiang Li, Ying Wu, and Gang Hua. Flexible visual recognition by evidential modeling of con- fusion and ignorance. In2023 IEEE/CVF International Conference on Computer Vision (ICCV), pages 1338–1347,

    work page

  7. [7]

    Evidential active recognition: Intelligent and pru- dent open-world embodied perception

    Lei Fan, Mingfu Liang, Yunxuan Li, Gang Hua, and Ying Wu. Evidential active recognition: Intelligent and pru- dent open-world embodied perception. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16351–16361, 2024. 3

    work page 2024

  8. [8]

    Towards discovery and attribution of open-world gan generated images.2021 IEEE/CVF In- ternational Conference on Computer Vision (ICCV), pages 14074–14083, 2021

    Sharath Girish, Saksham Suri, Saketh Rambhatla, and Ab- hinav Shrivastava. Towards discovery and attribution of open-world gan generated images.2021 IEEE/CVF In- ternational Conference on Computer Vision (ICCV), pages 14074–14083, 2021. 2

    work page 2021

  9. [9]

    Exploiting fine-grained face forgery clues via progressive enhancement learning

    Qiqi Gu, Shen Chen, Taiping Yao, Yang Chen, Shouhong Ding, and Ran Yi. Exploiting fine-grained face forgery clues via progressive enhancement learning. InProceedings of the AAAI Conference on Artificial Intelligence, pages 735–743,

    work page

  10. [10]

    Trufor: Leveraging all-round clues for trustworthy image forgery detection and localiza- tion

    Fabrizio Guillaro, Davide Cozzolino, Avneesh Sud, Nicholas Dufour, and Luisa Verdoliva. Trufor: Leveraging all-round clues for trustworthy image forgery detection and localiza- tion. InProceedings of the IEEE/CVF conference on com- puter vision and pattern recognition, pages 20606–20615,

    work page

  11. [11]

    Predictive uncertainty quantification of deep neural networks using dirichlet distributions

    Ahmed Hammam, Frank Bonarens, Seyed Eghbal Ghobadi, and Christoph Stiller. Predictive uncertainty quantification of deep neural networks using dirichlet distributions. InPro- ceedings of the 6th ACM Computer Science in Cars Sympo- sium, pages 1–10, 2022. 5

    work page 2022

  12. [12]

    Trusted multi-view classification

    Zongbo Han, Changqing Zhang, Huazhu Fu, and Joey Tianyi Zhou. Trusted multi-view classification. InInternational Conference on Learning Representations, 2020. 3, 5

    work page 2020

  13. [13]

    Trusted multi-view classification with dynamic evi- dential fusion.IEEE transactions on pattern analysis and machine intelligence, 45(2):2551–2566, 2022

    Zongbo Han, Changqing Zhang, Huazhu Fu, and Joey Tianyi Zhou. Trusted multi-view classification with dynamic evi- dential fusion.IEEE transactions on pattern analysis and machine intelligence, 45(2):2551–2566, 2022. 5

    work page 2022

  14. [14]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. InProceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 7

    work page 2016

  15. [15]

    Forgerynet: A versatile benchmark for comprehen- sive forgery analysis

    Yinan He, Bei Gan, Siyu Chen, Yichun Zhou, Guojun Yin, Luchuan Song, Lu Sheng, Jing Shao, and Ziwei Liu. Forgerynet: A versatile benchmark for comprehen- sive forgery analysis. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 4360–4369, 2021. 6

    work page 2021

  16. [16]

    Scaling out-of-distribution detection for real-world settings

    Dan Hendrycks, Steven Basart, Mantas Mazeika, Andy Zou, Joe Kwon, Mohammadreza Mostajabi, Jacob Steinhardt, and Dawn Song. Scaling out-of-distribution detection for real- world settings.arXiv preprint arXiv:1911.11132, 2019. 4

    work page arXiv 1911

  17. [17]

    Lora: Low-rank adaptation of large language models.ICLR, 1(2):3, 2022

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen- Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. Lora: Low-rank adaptation of large language models.ICLR, 1(2):3, 2022. 5

    work page 2022

  18. [18]

    To- wards generalized deepfake detection with continual learn- ing on limited new data

    He Huang, Nan Sun, Xufeng Lin, and Nour Moustafa. To- wards generalized deepfake detection with continual learn- ing on limited new data. In2022 International Conference on Digital Image Computing: Techniques and Applications (DICTA), pages 1–7. IEEE, 2022. 2

    work page 2022

  19. [19]

    Crest: Cross-modal resonance through evidential deep learn- ing for enhanced zero-shot learning

    Haojian Huang, Xiaozhennn Qiao, Zhuo Chen, Haodong Chen, Bingyu Li, Zhe Sun, Mulin Chen, and Xuelong Li. Crest: Cross-modal resonance through evidential deep learn- ing for enhanced zero-shot learning. InProceedings of the 32nd ACM International Conference on Multimedia, pages 5181–5190, 2024. 3

    work page 2024

  20. [20]

    What can discriminator do? towards box-free ownership verifica- tion of generative adversarial networks

    Ziheng Huang, Boheng Li, Yan Cai, Run Wang, Shang- wei Guo, Liming Fang, Jing Chen, and Lina Wang. What can discriminator do? towards box-free ownership verifica- tion of generative adversarial networks. InProceedings of the IEEE/CVF international conference on computer vision, pages 5009–5019, 2023. 2

    work page 2023

  21. [21]

    Springer, 2016

    Audun Jøsang.Subjective logic. Springer, 2016. 4

    work page 2016

  22. [22]

    Progressive Growing of GANs for Improved Quality, Stability, and Variation

    Tero Karras. Progressive growing of gans for improved qual- ity, stability, and variation.arXiv preprint arXiv:1710.10196,

    work page internal anchor Pith review Pith/arXiv arXiv

  23. [23]

    Oc-fakedect: Classifying deepfakes using one-class variational autoencoder

    Hasam Khalid and Simon S Woo. Oc-fakedect: Classifying deepfakes using one-class variational autoencoder. InPro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 656–657, 2020. 2, 3, 6, 7

    work page 2020

  24. [24]

    Clipping the deception: Adapting vision-language models for univer- sal deepfake detection

    Sohail Ahmed Khan and Duc-Tien Dang-Nguyen. Clipping the deception: Adapting vision-language models for univer- sal deepfake detection. InProceedings of the 2024 Inter- national Conference on Multimedia Retrieval, pages 1006– 1015, 2024. 2, 6, 7, 9

    work page 2024

  25. [25]

    Auto-Encoding Variational Bayes

    Diederik P Kingma and Max Welling. Auto-encoding varia- tional bayes.arXiv preprint arXiv:1312.6114, 2013. 7

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  26. [26]

    Fast face-swap using convolutional neural networks

    Iryna Korshunova, Wenzhe Shi, Joni Dambre, and Lucas Theis. Fast face-swap using convolutional neural networks. 10 InProceedings of the IEEE international conference on com- puter vision, pages 3677–3685, 2017. 1

    work page 2017

  27. [27]

    From coarse to fine-grained open-set recognition

    Nico Lang, V ´esteinn Snæbjarnarson, Elijah Cole, Oisin Mac Aodha, Christian Igel, and Serge Belongie. From coarse to fine-grained open-set recognition. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 17804–17814, 2024. 3

    work page 2024

  28. [28]

    Seeable: Soft discrepancies and bounded contrastive learning for exposing deepfakes

    Nicolas Larue, Ngoc-Son Vu, Vitomir Struc, Peter Peer, and Vassilis Christophides. Seeable: Soft discrepancies and bounded contrastive learning for exposing deepfakes. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 21011–21021, 2023. 2

    work page 2023

  29. [29]

    Cfpl- fas: Class free prompt learning for generalizable face anti- spoofing

    Ajian Liu, Shuai Xue, Jianwen Gan, Jun Wan, Yanyan Liang, Jiankang Deng, Sergio Escalera, and Zhen Lei. Cfpl- fas: Class free prompt learning for generalizable face anti- spoofing. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 222–232,

    work page

  30. [30]

    Spatial- phase shallow learning: rethinking face forgery detection in frequency domain

    Honggu Liu, Xiaodan Li, Wenbo Zhou, Yuefeng Chen, Yuan He, Hui Xue, Weiming Zhang, and Nenghai Yu. Spatial- phase shallow learning: rethinking face forgery detection in frequency domain. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 772–781, 2021. 6, 7

    work page 2021

  31. [31]

    Forgery-aware adaptive transformer for generalizable synthetic image detection

    Huan Liu, Zichang Tan, Chuangchuang Tan, Yunchao Wei, Jingdong Wang, and Yao Zhao. Forgery-aware adaptive transformer for generalizable synthetic image detection. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 10770–10780, 2024. 2

    work page 2024

  32. [32]

    Gener- alizing face forgery detection with high-frequency features

    Yuchen Luo, Yong Zhang, Junchi Yan, and Wei Liu. Gener- alizing face forgery detection with high-frequency features. InProceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 16317–16326, 2021. 2

    work page 2021

  33. [33]

    The creation and detection of deepfakes: A survey.ACM computing surveys (CSUR), 54(1):1–41, 2021

    Yisroel Mirsky and Wenke Lee. The creation and detection of deepfakes: A survey.ACM computing surveys (CSUR), 54(1):1–41, 2021. 1, 4

    work page 2021

  34. [34]

    On improving cross-dataset generalization of deepfake detectors

    Aakash Varma Nadimpalli and Ajita Rattani. On improving cross-dataset generalization of deepfake detectors. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 91–99,

    work page

  35. [35]

    Core: Consistent repre- sentation learning for face forgery detection

    Yunsheng Ni, Depu Meng, Changqian Yu, Chengbin Quan, Dongchun Ren, and Youjian Zhao. Core: Consistent repre- sentation learning for face forgery detection. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12–21, 2022. 1, 2

    work page 2022

  36. [36]

    Fsgan: Subject agnostic face swapping and reenactment

    Yuval Nirkin, Yosi Keller, and Tal Hassner. Fsgan: Subject agnostic face swapping and reenactment. InProceedings of the IEEE/CVF international conference on computer vision, pages 7184–7193, 2019. 1

    work page 2019

  37. [37]

    Towards uni- versal fake image detectors that generalize across genera- tive models

    Utkarsh Ojha, Yuheng Li, and Yong Jae Lee. Towards uni- versal fake image detectors that generalize across genera- tive models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24480– 24489, 2023. 2, 3, 6, 7, 9

    work page 2023

  38. [38]

    Advancing open-set domain generalization using evi- dential bi-level hardest domain scheduler.Advances in Neu- ral Information Processing Systems, 37:85412–85440, 2025

    Kunyu Peng, Di Wen, Kailun Yang, Ao Luo, Yufan Chen, Jia Fu, M Saquib Sarfraz, Alina Roitberg, and Rainer Stiefelha- gen. Advancing open-set domain generalization using evi- dential bi-level hardest domain scheduler.Advances in Neu- ral Information Processing Systems, 37:85412–85440, 2025. 4

    work page 2025

  39. [39]

    Thinking in frequency: Face forgery detection by min- ing frequency-aware clues

    Yuyang Qian, Guojun Yin, Lu Sheng, Zixuan Chen, and Jing Shao. Thinking in frequency: Face forgery detection by min- ing frequency-aware clues. InEuropean conference on com- puter vision, pages 86–103. Springer, 2020. 1

    work page 2020

  40. [40]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. InInternational conference on machine learning, pages 8748–8763. PMLR, 2021. 2, 4, 6, 7

    work page 2021

  41. [41]

    Faceforen- sics++: Learning to detect manipulated facial images

    Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- tian Riess, Justus Thies, and Matthias Nießner. Faceforen- sics++: Learning to detect manipulated facial images. In Proceedings of the IEEE/CVF international conference on computer vision, pages 1–11, 2019. 6, 7

    work page 2019

  42. [42]

    Toward open set recogni- tion.IEEE transactions on pattern analysis and machine intelligence, 35(7):1757–1772, 2012

    Walter J Scheirer, Anderson de Rezende Rocha, Archana Sapkota, and Terrance E Boult. Toward open set recogni- tion.IEEE transactions on pattern analysis and machine intelligence, 35(7):1757–1772, 2012. 3

    work page 2012

  43. [43]

    Eviden- tial deep learning to quantify classification uncertainty.Ad- vances in neural information processing systems, 31, 2018

    Murat Sensoy, Lance Kaplan, and Melih Kandemir. Eviden- tial deep learning to quantify classification uncertainty.Ad- vances in neural information processing systems, 31, 2018. 2, 3, 5

    work page 2018

  44. [44]

    Uncertainty-aware deep classifiers using generative models

    Murat Sensoy, Lance Kaplan, Federico Cerutti, and Maryam Saleki. Uncertainty-aware deep classifiers using generative models. InProceedings of the AAAI conference on artificial intelligence, pages 5620–5627, 2020. 2, 3

    work page 2020

  45. [45]

    Combination of evidence in dempster-shafer theory

    Kari Sentz and Scott Ferson. Combination of evidence in dempster-shafer theory. 2002. 2, 4, 5

    work page 2002

  46. [46]

    Learning residual images for face attribute manipulation

    Wei Shen and Rujie Liu. Learning residual images for face attribute manipulation. InProceedings of the IEEE con- ference on computer vision and pattern recognition, pages 4030–4038, 2017. 1

    work page 2017

  47. [47]

    Multi- faceted uncertainty estimation for label-efficient deep learn- ing.Advances in neural information processing systems, 33: 17247–17257, 2020

    Weishi Shi, Xujiang Zhao, Feng Chen, and Qi Yu. Multi- faceted uncertainty estimation for label-efficient deep learn- ing.Advances in neural information processing systems, 33: 17247–17257, 2020. 2, 3

    work page 2020

  48. [48]

    Detecting deep- fakes with self-blended images

    Kaede Shiohara and Toshihiko Yamasaki. Detecting deep- fakes with self-blended images. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18720–18729, 2022. 2, 3, 6, 7

    work page 2022

  49. [49]

    First order motion model for image animation.Advances in neural information processing systems, 32, 2019

    Aliaksandr Siarohin, St ´ephane Lathuili`ere, Sergey Tulyakov, Elisa Ricci, and Nicu Sebe. First order motion model for image animation.Advances in neural information processing systems, 32, 2019. 1

    work page 2019

  50. [50]

    An information theo- retic approach for attention-driven face forgery detection

    Ke Sun, Hong Liu, Taiping Yao, Xiaoshuai Sun, Shen Chen, Shouhong Ding, and Rongrong Ji. An information theo- retic approach for attention-driven face forgery detection. In European Conference on Computer Vision, pages 111–127. Springer, 2022. 2, 6, 7

    work page 2022

  51. [51]

    Contrastive pseudo learning for open-world deepfake attribution

    Zhimin Sun, Shen Chen, Taiping Yao, Bangjie Yin, Ran Yi, Shouhong Ding, and Lizhuang Ma. Contrastive pseudo learning for open-world deepfake attribution. InProceedings 11 of the IEEE/CVF International Conference on Computer Vi- sion, pages 20882–20892, 2023. 1, 2

    work page 2023

  52. [52]

    Rethinking the up-sampling op- erations in cnn-based generative network for generalizable deepfake detection

    Chuangchuang Tan, Yao Zhao, Shikui Wei, Guanghua Gu, Ping Liu, and Yunchao Wei. Rethinking the up-sampling op- erations in cnn-based generative network for generalizable deepfake detection. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 28130–28139, 2024. 2, 6, 7

    work page 2024

  53. [53]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. InInternational conference on machine learning, pages 6105–6114. PMLR,

    work page

  54. [54]

    Vim: Out-of-distribution with virtual-logit matching

    Haoqi Wang, Zhizhong Li, Litong Feng, and Wayne Zhang. Vim: Out-of-distribution with virtual-logit matching. InPro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4921–4930, 2022. 4

    work page 2022

  55. [55]

    Open set classification of gan-based image manipula- tions via a vit-based hybrid architecture

    Jun Wang, Omran Alamayreh, Benedetta Tondi, and Mauro Barni. Open set classification of gan-based image manipula- tions via a vit-based hybrid architecture. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 953–962, 2023. 2

    work page 2023

  56. [56]

    Bosc: A backdoor-based framework for open set synthetic image at- tribution.arXiv preprint arXiv:2405.11491, 2024

    Jun Wang, Benedetta Tondi, and Mauro Barni. Bosc: A backdoor-based framework for open set synthetic image at- tribution.arXiv preprint arXiv:2405.11491, 2024. 2

    work page arXiv 2024

  57. [57]

    Towards evidential and class separable open set object de- tection

    Ruofan Wang, Rui-Wei Zhao, Xiaobo Zhang, and Rui Feng. Towards evidential and class separable open set object de- tection. InProceedings of the AAAI Conference on Artificial Intelligence, pages 5572–5580, 2024. 4

    work page 2024

  58. [58]

    Cnn-generated images are surprisingly easy to spot

    Sheng-Yu Wang, Oliver Wang, Richard Zhang, Andrew Owens, and Alexei A Efros. Cnn-generated images are surprisingly easy to spot... for now. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8695–8704, 2020. 1, 2

    work page 2020

  59. [59]

    Energy-based open-world uncertainty modeling for confidence calibration

    Yezhen Wang, Bo Li, Tong Che, Kaiyang Zhou, Ziwei Liu, and Dongsheng Li. Energy-based open-world uncertainty modeling for confidence calibration. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 9302–9311, 2021. 3

    work page 2021

  60. [60]

    Dire for diffusion-generated image detection

    Zhendong Wang, Jianmin Bao, Wengang Zhou, Weilun Wang, Hezhen Hu, Hong Chen, and Houqiang Li. Dire for diffusion-generated image detection. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 22445–22455, 2023. 2

    work page 2023

  61. [61]

    Traceevader: Making deepfakes more untraceable via evading the forgery model attribution

    Mengjie Wu, Jingui Ma, Run Wang, Sidan Zhang, Ziyou Liang, Boheng Li, Chenhao Lin, Liming Fang, and Lina Wang. Traceevader: Making deepfakes more untraceable via evading the forgery model attribution. InProceedings of the AAAI Conference on Artificial Intelligence, pages 19965– 19973, 2024. 2

    work page 2024

  62. [62]

    Ucf: Uncovering common features for generalizable deep- fake detection

    Zhiyuan Yan, Yong Zhang, Yanbo Fan, and Baoyuan Wu. Ucf: Uncovering common features for generalizable deep- fake detection. InProceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, pages 22412–22423,

    work page

  63. [63]

    Deepfakebench: A comprehensive benchmark of deepfake detection.arXiv preprint arXiv:2307.01426,

    Zhiyuan Yan, Yong Zhang, Xinhang Yuan, Siwei Lyu, and Baoyuan Wu. Deepfakebench: A comprehensive benchmark of deepfake detection.arXiv preprint arXiv:2307.01426,

    work page arXiv

  64. [64]

    Transcending forgery specificity with latent space augmentation for generalizable deepfake detection

    Zhiyuan Yan, Yuhao Luo, Siwei Lyu, Qingshan Liu, and Baoyuan Wu. Transcending forgery specificity with latent space augmentation for generalizable deepfake detection. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 8984–8994, 2024. 3

    work page 2024

  65. [65]

    DF40: Toward next-generation deepfake detection

    Zhiyuan Yan, Taiping Yao, Shen Chen, Yandan Zhao, Xinghe Fu, Junwei Zhu, Donghao Luo, Chengjie Wang, Shouhong Ding, Yunsheng Wu, et al. Df40: To- ward next-generation deepfake detection.arXiv preprint arXiv:2406.13495, 2024. 6

    work page arXiv 2024

  66. [66]

    Convolutional prototype network for open set recognition.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(5):2358–2370, 2020

    Hong-Ming Yang, Xu-Yao Zhang, Fei Yin, Qing Yang, and Cheng-Lin Liu. Convolutional prototype network for open set recognition.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(5):2358–2370, 2020. 3

    work page 2020

  67. [67]

    Deepfake network architecture attribution

    Tianyun Yang, Ziyao Huang, Juan Cao, Lei Li, and Xirong Li. Deepfake network architecture attribution. InProceed- ings of the AAAI Conference on Artificial Intelligence, pages 4662–4670, 2022. 2

    work page 2022

  68. [68]

    Dˆ 3: Scaling up deepfake detection by learning from discrepancy

    Yongqi Yang, Zhihao Qian, Ye Zhu, Olga Russakovsky, and Yu Wu. Dˆ 3: Scaling up deepfake detection by learning from discrepancy. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 23850–23859,

    work page

  69. [69]

    Anedl: adap- tive negative evidential deep learning for open-set semi- supervised learning

    Yang Yu, Danruo Deng, Furui Liu, Qi Dou, Yueming Jin, Guangyong Chen, and Pheng Ann Heng. Anedl: adap- tive negative evidential deep learning for open-set semi- supervised learning. InProceedings of the AAAI Conference on Artificial Intelligence, pages 16587–16595, 2024. 4

    work page 2024

  70. [70]

    Prototypical matching and open set rejection for zero-shot semantic segmentation

    Hui Zhang and Henghui Ding. Prototypical matching and open set rejection for zero-shot semantic segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 6974–6983, 2021. 3

    work page 2021

  71. [71]

    Detecting and simulating artifacts in gan fake images

    Xu Zhang, Svebor Karaman, and Shih-Fu Chang. Detecting and simulating artifacts in gan fake images. In2019 IEEE in- ternational workshop on information forensics and security (WIFS), pages 1–6. IEEE, 2019. 2

    work page 2019

  72. [72]

    Decoupling maxlogit for out- of-distribution detection

    Zihan Zhang and Xiang Xiang. Decoupling maxlogit for out- of-distribution detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3388–3397, 2023. 3, 7

    work page 2023

  73. [73]

    Open set action recognition via multi-label eviden- tial learning

    Chen Zhao, Dawei Du, Anthony Hoogs, and Christopher Funk. Open set action recognition via multi-label eviden- tial learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 22982–22991, 2023. 4

    work page 2023

  74. [74]

    Copyright protection and accountability of generative ai: Attack, watermarking and at- tribution

    Haonan Zhong, Jiamin Chang, Ziyue Yang, Tingmin Wu, Pathum Chamikara Mahawaga Arachchige, Chehara Pathmabandu, and Minhui Xue. Copyright protection and accountability of generative ai: Attack, watermarking and at- tribution. InCompanion Proceedings of the ACM Web Con- ference 2023, pages 94–98, 2023. 2

    work page 2023

  75. [75]

    Learning to prompt for vision-language models.In- ternational Journal of Computer Vision, 130(9):2337–2348,

    Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. Learning to prompt for vision-language models.In- ternational Journal of Computer Vision, 130(9):2337–2348,

    work page

  76. [76]

    Fine- grained open-set deepfake detection via unsupervised do- 12 main adaptation.IEEE Transactions on Information Foren- sics and Security, 2024

    Xinye Zhou, Hu Han, Shiguang Shan, and Xilin Chen. Fine- grained open-set deepfake detection via unsupervised do- 12 main adaptation.IEEE Transactions on Information Foren- sics and Security, 2024. 1

    work page 2024

  77. [77]

    Uia- vit: Unsupervised inconsistency-aware method based on vi- sion transformer for face forgery detection

    Wanyi Zhuang, Qi Chu, Zhentao Tan, Qiankun Liu, Haojie Yuan, Changtao Miao, Zixiang Luo, and Nenghai Yu. Uia- vit: Unsupervised inconsistency-aware method based on vi- sion transformer for face forgery detection. InEuropean con- ference on computer vision, pages 391–407. Springer, 2022. 1, 2 13

    work page 2022