Pith. sign in

REVIEW 3 major objections 6 minor 53 references

Stacking Brick by Brick: Aligned Feature Isolation for Incremental Face Forgery Detection

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Incremental face forgery detectors can avoid catastrophic forgetting by isolating each task's feature distribution in latent space and aligning their decision boundaries, the paper claims.

desk verdict Solid, well-engineered IFFD paper with a new replay strategy and feature-isolation pipeline; the gains look real but need error bars and a boundary-alignment caveat. read the letter →

arxiv 2411.11396 v3 pith:5BL7VYA5 submitted 2024-11-18 cs.CV

Jikang Cheng , Zhiyuan Yan , Ying Zhang , Li Hao , Jiaxin Ai , Qin Zou , Chen Li , Zhongyuan Wang This is my paper · ORCID
classification cs.CV
keywords incrementallearningfaceforgerydetectiondeepfakecatastrophicforgettingfeatureisolationreplaystrategycontrastivelatentspace
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that incremental face forgery detectors forget old forgeries because treating every forgery as one 'fake' class lets the new task's feature distribution overwrite the previous ones. It proposes to instead keep each task's distribution as a separate, isolated cluster in latent space, aligned so that the Real/Fake boundary stays consistent across tasks. The claim is that this 'brick-by-brick' stacking preserves old forgery knowledge while accumulating new forgery types, and the reported experiments support it with higher AUC than previous replay-based methods on two newly constructed protocols. A sympathetic reader would care because it offers a concrete way to update deepfake detectors continuously without retraining on all past data.

What carries the argument

Two components carry the argument. Sparse Uniform Replay (SUR) selects, within each magnitude segment, the most shuffle-stable feature and a feature with the lowest angular similarity, producing a uniformly sparse version of the previous global distribution. The Latent-space Incremental Detector (LID) uses these replayed points plus distribution re-filling—mixing between replay features and their centroid—to feed a supervised contrastive isolation loss, while Incremental Decision Alignment recursively rotates the new task's linear classifier weight vector toward the previous one so all per-task decision boundaries align.

What would settle it

Compute, for each replayed image, the distance between its feature under the frozen old encoder and its feature under the updated encoder after each increment. If that drift is large for many replay samples, the old task's 'brick' is built from stale points; a direct experiment is to compare Protocol 1 performance when replay features are recomputed on the old backbone versus the current backbone, or when replay images are re-selected each step.

Watch

Extended reading notes

Core claim

The central discovery is that replay-based incremental face forgery detection works better when the replay set is chosen to approximate the whole previous distribution—not just center or hard samples—and when training explicitly separates and aligns per-task distributions. The paper reports that its Sparse Uniform Replay (SUR) selection, which picks uniformly sparse, stable samples by magnitude, angularity, and shuffle-consistency, combined with a contrastive isolation loss and decision-boundary alignment, raises average AUC from 0.8752 to 0.9315 on Protocol 1 and from 0.8141 to 0.9433 on Protocol 2 compared with HDP.

Load-bearing premise

The load-bearing premise is that the SUR replay set, selected with the frozen backbone, still resembles the old task's global feature distribution after the backbone is fine-tuned on the new task, since all isolation and alignment losses depend on those replayed points.

Editorial extensions

If this is right

  • Per-task feature isolation makes earlier-task accuracy degrade far less when new forgery types arrive, because old clusters are not overwritten.
  • Because each task's real and fake clusters stay separate, the detector accumulates diverse forgery cues instead of compressing them into a single 'fake' blob.
  • The aligned decision boundaries let the final detector average all task classifiers at inference, so it can exploit accumulated knowledge even when the task identity is unknown.
  • The approach also improves generalization to unseen datasets and to other backbone architectures, indicating the accumulated distributions carry generalizable forgery information.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension would be to apply the SUR-LID recipe to other binary forgery or anomaly detection settings where the negative class is method-specific rather than semantically homogeneous.
  • The method's fixed replay budget implies a testable tradeoff: as the number of tasks grows, the per-task replay budget shrinks, and at some point distribution re-filling may no longer compensate; measuring that breakpoint would bound the method's scalability.
  • The shuffle-stability criterion assumes forgery features are invariant to grid shuffling; forgeries that alter global layout rather than local texture could be mis-ranked by this criterion, which is a testable extension.
  • One could combine SUR with generative replay to refresh old distributions, directly addressing the stale-replay drift that the paper does not measure.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper proposes SUR-LID, an incremental face forgery detection method built on two components: Sparse Uniform Replay (SUR), which selects replay samples that approximate the previous task's global feature distribution, and a Latent-space Incremental Detector (LID), which adds isolation and decision-alignment losses so that each new task's feature distribution is stacked without overriding previous ones. The authors introduce two incremental protocols (dataset-incremental and forgery-category-incremental) and report large AUC gains over prior replay-based IFFD methods, together with ablations, UMAP/Grad-CAM visualizations, generalization experiments, and robustness checks. The core claim is that aligned feature isolation mitigates catastrophic forgetting and accumulates diverse forgery information.

Significance. If the gains are reproducible, the paper would be a meaningful advance for continual deepfake detection: the proposed protocols are more realistic than the single classical protocol used in earlier work, and the reported margins on Protocol 2 are substantial. The work also has practical strengths: code is released, ablations isolate the contribution of SUR, DR, Liso, and IDA, and the authors include cross-dataset and robustness evaluations. However, the central mechanism rests on an unverified invariance assumption about replay-set representativeness after backbone updates, the decision-alignment step does not actually align full linear decision boundaries, and the main results are reported without variance, so the significance of the numerical margins is not established.

major comments (3)
  1. [Section 3.4.2 / Eq. (6)] The claim that aligning decision boundaries is identical to ensuring angularity consistency of the linear parameters is incomplete. For a linear classifier C_t(f) = theta_t * f + b_t, the decision boundary is theta_t * f + b_t = 0; Eq. (6) updates only the weight direction theta_{t+1} and says nothing about the bias b_{t+1}. Aligning weight directions makes the task-specific hyperplanes parallel, not coincident, and the inference-time averaging in Eq. (10) then combines classifiers with potentially different offsets. Please add an explicit bias-alignment mechanism, or provide evidence that bias differences are negligible in practice.
  2. [Sections 3.3-3.4 / Eq. (7)] SUR selects replay samples using features from the frozen encoder E_t, but during task t+1 training the replayed images are re-embedded with the updated encoder E_{t+1}, and Eqs. (4), (5), and (8) operate on these new features. The only direct constraint on the old feature space is the distillation loss L_dis in Eq. (7), which anchors the replayed points but does not constrain the rest of the old manifold. If E_{t+1} drifts or rotates the previous feature distribution non-uniformly, the replay set is no longer a uniformly sparse proxy of the old global distribution, so the distribution re-filling in Eq. (4) is built from drifted landmarks. The MMD evidence in Fig. 5 is computed at selection time with E_t, not after subsequent increments. Please report a per-increment drift statistic, such as MMD between E_{t+1}(replay) and E_t(original training features) or per-domain feature shift, to substantiate the central stacking assumption.
  3. [Tables 1-3 and Supplementary Table 5] No variance, standard deviation, or number of independent runs is reported for any AUC value. The main comparisons and ablations therefore do not establish that the reported differences are statistically significant, especially for per-cell margins that are a few points (e.g., FF++ under Protocol 1). Please report mean +/- std over at least three seeds for the main tables and, where feasible, paired significance tests for the comparisons against prior methods and for the ablation variants.
minor comments (6)
  1. [Table 1] The iCaRL row has arithmetic inconsistencies in the Protocol 1 averages: at T2, 0.9267 and 0.9479 average to 0.9373, not 0.8363, and at T3, 0.9010, 0.7447, and 0.9135 average to 0.8531, not 0.7864. Please correct these entries or explain the reported averages.
  2. [Eq. (5)] The isolation loss denominator sums only over k with y_i != y_k, whereas the standard supervised contrastive denominator includes all other samples, including positive pairs. If the omission is intentional, please state this explicitly; otherwise the formula should be corrected.
  3. [Section 4.1 and Eq. (5)] The implementation details list mu1, mu2, and gamma but do not report the value of the contrastive temperature tau used in Eq. (5).
  4. [Eq. (6)] The displayed update for theta_{t+1} is typeset in a garbled way; please rewrite it with clear normalization and interpolation steps so that the intended operation is unambiguous.
  5. [Section 4.3] There is a typo in the ablation subsection title: 'Sprase Uniform Replay' should read 'Sparse Uniform Replay'.
  6. [Supplementary Table 4] The Protocol 3 results for previous methods are copied from other papers under different experimental settings, as the text acknowledges; this should be labeled as not directly comparable rather than presented as a head-to-head comparison, or the baselines should be reproduced under the same protocol.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: SUR-LID is an empirical pipeline tested on external benchmarks; minor self-citations are not load-bearing.

full rationale

The derivation is self-contained and empirically grounded. SUR's replay selection (Eqs. 1-3) uses the current backbone's feature geometry to choose stable, uniform samples; this is a design heuristic, not a quantity fitted to the evaluation target. LID's isolation loss (Eq. 5), distribution re-filling (Eq. 4), decision alignment (Eq. 6), and distillation (Eq. 7) all operate on publicly defined losses and constraints; no loss term encodes the benchmark AUCs or the test labels. The headline results (Tab. 1) are measured on held-out splits of public datasets (SDv21, FF++, DFDCP, CDF, DF40 subsets) against independently reproduced baselines, so the benchmark does not reduce to the method's inputs. The paper's self-citations [4,5] are used for shuffle consistency and for the general claim that latent-space organization matters, but these uses are corroborated by external citations ([29,38] and [7,11]) and are not load-bearing for the central isolation/alignment claim. The main caveat, explicitly admitted in Supp. Sec. 5.1, is that at small replay sizes "the constraints employed for the proposed aligned feature isolation rely heavily on the replayed global distribution"; and the MMD evidence in Fig. 5 is measured with the selection-time backbone, so representativeness after later fine-tuning is not directly verified. These are correctness and robustness risks, not circularity.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the replay set remaining a faithful proxy of old distributions, on the shuffle-stability criterion capturing forgery-relevant features, and on the average of aligned classifiers being a valid decision function. The free parameters are hyperparameters chosen for the reported configuration.

free parameters (5)
  • Replay buffer size per task (nr) = 500
    Chosen by hand. The sensitivity study in Supplementary 5.1 shows performance degrades at smaller sizes, so the reported results depend on this value.
  • Trade-off weight for distillation loss (µ1) = 1
    Set in Section 4.1; controls how strongly old-feature consistency is enforced, which underpins the replay representativeness assumption.
  • Trade-off weight for detection loss (µ2) = 0.1
    Set in Section 4.1; balances final binary detection with isolation and distillation terms.
  • Alignment learning rate (γ) = 0.001
    Set in Section 4.1; determines how aggressively the new task classifier is rotated toward previous classifiers in Eq. 6.
  • Contrastive temperature (τ) = not stated
    Used in Eq. 5 for isolation loss; value is not reported, which is a reproducibility gap.
assumptions (3)
  • domain assumption Grid shuffle preserves fine-grained forgery clues; stable features under shuffle are forgery-relevant.
    Section 3.3, Eq. 3 applies shuffle consistency, citing prior works [5,29,38]. If this does not hold, replay selection drops relevant samples.
  • domain assumption A replay subset selected with old features remains representative of the previous global distribution after backbone updates.
    This is the weakest structural premise: SUR is computed with E_t, but used with E_{t+1} during incremental training (Sections 3.4, 3.5).
  • domain assumption Averaging aligned linear classifiers at inference yields a valid binary decision.
    Eq. 10 uses the average of per-task classifiers; this assumes their aligned boundaries are close enough that averaging preserves the real/fake separation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Stacking Brick by Brick: Aligned Feature Isolation for Incremental Face Forgery Detection." pith.science (2026). https://pith.science/paper/5BL7VYA5

@misc{pith2026241111396,
  author       = {Pith},
  title        = {Pith review of: Stacking Brick by Brick: Aligned Feature Isolation for Incremental Face Forgery Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5BL7VYA5}},
  note         = {Machine review of arXiv:2411.11396}
}
abstract

The rapid advancement of face forgery techniques has introduced a growing variety of forgeries. Incremental Face Forgery Detection (IFFD), involving gradually adding new forgery data to fine-tune the previously trained model, has been introduced as a promising strategy to deal with evolving forgery methods. However, a naively trained IFFD model is prone to catastrophic forgetting when new forgeries are integrated, as treating all forgeries as a single ''Fake" class in the Real/Fake classification can cause different forgery types overriding one another, thereby resulting in the forgetting of unique characteristics from earlier tasks and limiting the model's effectiveness in learning forgery specificity and generality. In this paper, we propose to stack the latent feature distributions of previous and new tasks brick by brick, $\textit{i.e.}$, achieving $\textbf{aligned feature isolation}$. In this manner, we aim to preserve learned forgery information and accumulate new knowledge by minimizing distribution overriding, thereby mitigating catastrophic forgetting. To achieve this, we first introduce Sparse Uniform Replay (SUR) to obtain the representative subsets that could be treated as the uniformly sparse versions of the previous global distributions. We then propose a Latent-space Incremental Detector (LID) that leverages SUR data to isolate and align distributions. For evaluation, we construct a more advanced and comprehensive benchmark tailored for IFFD. The leading experimental results validate the superiority of our method.

Figures

Figures reproduced from arXiv: 2411.11396 by the authors.

Figure 1
Figure 1. Illustration of the proposed aligned feature isolation in the [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overall framework of the proposed method. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Illustration of different replay strategies. Using Cen [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: UMAP [27] latent-space visualization for IFFD with Protocol 1. The upper row is the results of the baseline method (DFIL [30]) while the lower row is Ours. All shapes in blue are added for better illustration. The dashed lines denote the aligned boundary that divides r…
Figure 5
Figure 5. Figure 5: Evaluations of global distinction between the replay set [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Saliency map visualization of DFIL [30] and the proposed method. Baseline Ours [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Actual two-dimensional feature distributions of toy models with Protocol 1. 3. Experiments of Generalization Ability 3.1. Generalization to Other Unseen Datasets To validate that the accumulated forgery information en￾ables our method to learn more about forgery genera…
Figure 8
Figure 8. Figure 8: Sensitivity of replay size. The shown AUCs are the [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Robustness evaluations. The images in the first column are visualized illustrations of different types of applied perturbations. The [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 44 canonical work pages

  1. [1]

    Memory aware synapses: Learning what (not) to forget

    Rahaf Aljundi, Francesca Babiloni, Mohamed Elhoseiny, Marcus Rohrbach, and Tinne Tuytelaars. Memory aware synapses: Learning what (not) to forget. In ECCV, pages 139–154, 2018

    work page 2018

  2. [2]

    Destruction and construction learning for fine-grained image recognition

    Yue Chen, Yalong Bai, Wei Zhang, and Tao Mei. Destruction and construction learning for fine-grained image recognition. In CVPR, pages 5157–5166, 2019

    work page 2019

  3. [3]

    Diffusionface: Towards a comprehensive dataset for diffusion-based face forgery anal- ysis

    Zhongxi Chen, Ke Sun, Ziyin Zhou, Xianming Lin, Xiaoshuai Sun, Liujuan Cao, and Rongrong Ji. Diffusionface: Towards a comprehensive dataset for diffusion-based face forgery anal- ysis. arXiv preprint arXiv:2403.18471, 2024

    arXiv 2024

  4. [4]

    Can we leave deepfake data behind in training deepfake detector? arXiv preprint arXiv:2408.17052, 2024

    Jikang Cheng, Zhiyuan Yan, Ying Zhang, Yuhao Luo, Zhongyuan Wang, and Chen Li. Can we leave deepfake data behind in training deepfake detector? arXiv preprint arXiv:2408.17052, 2024

    arXiv 2024

  5. [5]

    ED$^4$: Explicit Data-level Debiasing for Deepfake Detection

    Jikang Cheng, Ying Zhang, Qin Zou, Zhiyuan Yan, Chao Liang, Zhongyuan Wang, and Chen Li. Ed 4: Explicit data-level debiasing for deepfake detection. arXiv preprint arXiv:2408.06779, 2024

    work page Pith review arXiv 2024

  6. [6]

    Xception: Deep learning with depthwise separable convolutions

    Franc ¸ois Chollet. Xception: Deep learning with depthwise separable convolutions. In CVPR, pages 1251–1258, 2017

    work page 2017

  7. [7]

    Group equivariant convolu- tional networks

    Taco Cohen and Max Welling. Group equivariant convolu- tional networks. In ICML, pages 2990–2999. PMLR, 2016

    work page 2016

  8. [8]

    A continual learning survey: Defying forgetting in classification tasks

    Matthias De Lange, Rahaf Aljundi, Marc Masana, Sarah Parisot, Xu Jia, Aleˇs Leonardis, Gregory Slabaugh, and Tinne Tuytelaars. A continual learning survey: Defying forgetting in classification tasks. IEEE TPAMI, 44(7):3366–3385, 2021

    work page 2021

Show all 53 references
  1. [9]

    https://www.kaggle

    Deepfake detection challenge. https://www.kaggle. com / c / deepfake - detection - challenge Ac- cessed 2021-04-24

    2021

  2. [10]

    https://ai.googleblog.com/2019/09/ contributing-data-to-deepfake-detection

    DFD. https://ai.googleblog.com/2019/09/ contributing-data-to-deepfake-detection. html Accessed 2021-04-24

    2019

  3. [11]

    To- wards deep symbolic reinforcement learning

    Marta Garnelo, Kai Arulkumaran, and Murray Shanahan. To- wards deep symbolic reinforcement learning. arXiv preprint arXiv:1609.05518, 2016

    2016 arXiv

  4. [12]

    A kernel two-sample test

    Arthur Gretton, Karsten M Borgwardt, Malte J Rasch, Bern- hard Sch¨olkopf, and Alexander Smola. A kernel two-sample test. JMLR, 13(1):723–773, 2012

    2012

  5. [13]

    Lips don’t lie: A generalisable and robust approach to face forgery detection

    Alexandros Haliassos, Konstantinos V ougioukas, Stavros Petridis, and Maja Pantic. Lips don’t lie: A generalisable and robust approach to face forgery detection. In CVPR, 2021

    2021

  6. [14]

    Implicit identity representation conditioned memory compensation network for talking head video generation

    Fa-Ting Hong and Dan Xu. Implicit identity representation conditioned memory compensation network for talking head video generation. In ICCV, pages 23062–23072, 2023

    2023

  7. [15]

    Implicit identity driven deepfake face swapping detection

    Baojin Huang, Zhongyuan Wang, Jifan Yang, Jiaxin Ai, Qin Zou, Qian Wang, and Dengpan Ye. Implicit identity driven deepfake face swapping detection. In CVPR, pages 4490– 4499, 2023

    2023

  8. [16]

    Alias-free generative adversarial networks

    Tero Karras, Miika Aittala, Samuli Laine, Erik H ¨ark¨onen, Janne Hellsten, Jaakko Lehtinen, and Timo Aila. Alias-free generative adversarial networks. NeurIPS, 34:852–863, 2021

    2021

  9. [17]

    Fakeavceleb: A novel audio-video multimodal deepfake dataset

    Hasam Khalid, Shahroz Tariq, Minha Kim, and Simon S Woo. Fakeavceleb: A novel audio-video multimodal deepfake dataset. arXiv preprint arXiv:2108.05080, 2021

    2021 arXiv

  10. [18]

    Supervised contrastive learning

    Prannay Khosla, Piotr Teterwak, Chen Wang, Aaron Sarna, Yonglong Tian, Phillip Isola, Aaron Maschinot, Ce Liu, and Dilip Krishnan. Supervised contrastive learning. NeurIPS, 33:18661–18673, 2020

    2020

  11. [19]

    Cored: Gen- eralizing fake media detection with continual representation using distillation

    Minha Kim, Shahroz Tariq, and Simon S Woo. Cored: Gen- eralizing fake media detection with continual representation using distillation. In ACM MM, pages 337–346, 2021

    2021

  12. [20]

    Overcoming catastrophic forgetting in neural networks

    James Kirkpatrick, Razvan Pascanu, Neil Rabinowitz, Joel Veness, Guillaume Desjardins, Andrei A Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, et al. Overcoming catastrophic forgetting in neural networks. Proceedings of the national academy of scien...

    2017

  13. [21]

    Celeb-df: A new dataset for deepfake forensics

    Yuezun Li, Xin Yang, Pu Sun, Honggang Qi, and Siwei Lyu. Celeb-df: A new dataset for deepfake forensics. In CVPR, 2020

    2020

  14. [22]

    Learning without forgetting

    Zhizhong Li and Derek Hoiem. Learning without forgetting. IEEE TPAMI, 40(12):2935–2947, 2017

    2017

  15. [23]

    Exploring disentangled content information for face forgery detection

    Jiahao Liang, Huafeng Shi, and Weihong Deng. Exploring disentangled content information for face forgery detection. In ECCV, pages 128–145. Springer, 2022

    2022

  16. [24]

    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. In CVPR, pages 772–781, 2021

    2021

  17. [25]

    Mnemonics training: Multi-class incremental learning without forgetting

    Yaoyao Liu, Yuting Su, An-An Liu, Bernt Schiele, and Qianru Sun. Mnemonics training: Multi-class incremental learning without forgetting. In CVPR, pages 12245–12254, 2020

    2020

  18. [26]

    Supervised contrastive replay: Revisiting the nearest class mean classifier in online class-incremental continual learning

    Zheda Mai, Ruiwen Li, Hyunwoo Kim, and Scott Sanner. Supervised contrastive replay: Revisiting the nearest class mean classifier in online class-incremental continual learning. In CVPR, pages 3589–3599, 2021

    2021

  19. [27]

    Umap: Uniform manifold approximation and projection for dimen- sion reduction

    Leland McInnes, John Healy, and James Melville. Umap: Uniform manifold approximation and projection for dimen- sion reduction. arXiv preprint arXiv:1802.03426, 2018

    2018 arXiv

  20. [28]

    Universal adversarial perturba- tions

    Seyed-Mohsen Moosavi-Dezfooli, Alhussein Fawzi, Omar Fawzi, and Pascal Frossard. Universal adversarial perturba- tions. In CVPR, pages 1765–1773, 2017

    2017

  21. [29]

    Core: Consistent represen- tation learning for face forgery detection

    Yunsheng Ni, Depu Meng, Changqian Yu, Chengbin Quan, Dongchun Ren, and Youjian Zhao. Core: Consistent represen- tation learning for face forgery detection. In CVPR Workshop, pages 12–21, 2022

    2022

  22. [30]

    Dfil: Deepfake incremental learning by exploiting domain- invariant forgery clues

    Kun Pan, Yifang Yin, Yao Wei, Feng Lin, Zhongjie Ba, Zhen- guang Liu, Zhibo Wang, Lorenzo Cavallaro, and Kui Ren. Dfil: Deepfake incremental learning by exploiting domain- invariant forgery clues. In ACM MM, pages 8035–8046, 2023

    2023

  23. [31]

    icarl: Incremental classifier and representation learning

    Sylvestre-Alvise Rebuffi, Alexander Kolesnikov, Georg Sperl, and Christoph H Lampert. icarl: Incremental classifier and representation learning. In CVPR, pages 2001–2010, 2017

    2001

  24. [32]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. CVPR, pages 10684– 10695, 2022

    2022

  25. [33]

    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 ICCV, pages 1–11, 2019

    2019

  26. [34]

    Grad- cam: Visual explanations from deep networks via gradient- based localization

    Ramprasaath R Selvaraju, Michael Cogswell, Abhishek Das, Ramakrishna Vedantam, Devi Parikh, and Dhruv Batra. Grad- cam: Visual explanations from deep networks via gradient- based localization. In ICCV, pages 618–626, 2017

    2017

  27. [35]

    Stable learning via differentiated variable decorrelation

    Zheyan Shen, Peng Cui, Jiashuo Liu, Tong Zhang, Bo Li, and Zhitang Chen. Stable learning via differentiated variable decorrelation. In KDD, pages 2185–2193, 2020

    2020

  28. [36]

    Detecting deep- fakes with self-blended images

    Kaede Shiohara and Toshihiko Yamasaki. Detecting deep- fakes with self-blended images. In CVPR, pages 18720– 18729, 2022

    2022

  29. [37]

    Blendface: Re-designing identity encoders for face-swapping

    Kaede Shiohara, Xingchao Yang, and Takafumi Taketomi. Blendface: Re-designing identity encoders for face-swapping. In ICCV, pages 7634–7644, 2023

    2023

  30. [38]

    Dual contrastive learning for general face forgery detection

    Ke Sun, Taiping Yao, Shen Chen, Shouhong Ding, Jilin Li, and Rongrong Ji. Dual contrastive learning for general face forgery detection. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 2316–2324, 2022

    2022

  31. [39]

    Continual face forgery detection via historical distribution preserving

    Ke Sun, Shen Chen, Taiping Yao, Xiaoshuai Sun, Shouhong Ding, and Rongrong Ji. Continual face forgery detection via historical distribution preserving. IJCV, pages 1–18, 2024

    2024

  32. [40]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In ICML, pages 6105–6114. PMLR, 2019

    2019

  33. [41]

    Dynamic mixed-prototype model for incremental deepfake detection

    Jiahe Tian, Cai Yu, Peng Chen, Zihao Xiao, Xi Wang, Jizhong Han, and Yesheng Chai. Dynamic mixed-prototype model for incremental deepfake detection. In ACM MM, 2024

    2024

  34. [42]

    Designing one unified framework for high-fidelity face reenactment and swapping

    Chao Xu, Jiangning Zhang, Yue Han, Guanzhong Tian, Xian- fang Zeng, Ying Tai, Yabiao Wang, Chengjie Wang, and Yong Liu. Designing one unified framework for high-fidelity face reenactment and swapping. In ECCV, pages 54–71. Springer, 2022

    2022

  35. [43]

    Der: Dynami- cally expandable representation for class incremental learning

    Shipeng Yan, Jiangwei Xie, and Xuming He. Der: Dynami- cally expandable representation for class incremental learning. In CVPR, pages 3014–3023, 2021

    2021

  36. [44]

    Ucf: Uncovering common features for generalizable deepfake detection

    Zhiyuan Yan, Yong Zhang, Yanbo Fan, and Baoyuan Wu. Ucf: Uncovering common features for generalizable deepfake detection. ICCV, 2023

    2023

  37. [45]

    Deepfakebench: A comprehensive benchmark of deepfake detection

    Zhiyuan Yan, Yong Zhang, Xinhang Yuan, Siwei Lyu, and Baoyuan Wu. Deepfakebench: A comprehensive benchmark of deepfake detection. arXiv preprint arXiv:2307.01426 , 2023

    2023 arXiv

  38. [46]

    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 CVPR, pages 8984–8994, 2024

    2024

  39. [47]

    Df40: Toward next-generation deep- fake detection

    Zhiyuan Yan, Taiping Yao, Shen Chen, Yandan Zhao, Xinghe Fu, Junwei Zhu, Donghao Luo, Li Yuan, Chengjie Wang, Shouhong Ding, et al. Df40: Toward next-generation deep- fake detection. arXiv preprint arXiv:2406.13495, 2024

    2024 arXiv

  40. [48]

    Deep stable learning for out-of-distribution generalization

    Xingxuan Zhang, Peng Cui, Renzhe Xu, Linjun Zhou, Yue He, and Zheyan Shen. Deep stable learning for out-of-distribution generalization. In CVPR, pages 5372–5382, 2021. Supplementary Materials

    2021

  41. [49]

    Results with Protocol 3 Method FF++ DFDCP DFD CDF Avg

    Further Results Comparing with SoTA 1.1. Results with Protocol 3 Method FF++ DFDCP DFD CDF Avg. LwF [22] 67.34 67.43 84.05 87.90 76.68 CoReD [19] 74.08 76.59 93.41 80.78 81.22 DFIL [30] 86.28 79.53 92.36 83.81 85.49 DMP [41] 91.61 84.86 91.81 91.67 89.99 Ours 90.89 89.33 93.97...

  42. [50]

    Visualization of Model Attention via Grad- CAM As shown in Fig 6, we deploy Grad-CAM [34] to generate saliency maps

    Further Visualization Analysis 2.1. Visualization of Model Attention via Grad- CAM As shown in Fig 6, we deploy Grad-CAM [34] to generate saliency maps. It can be observed that our method could ex- plore more forgery clues since we successfully accumulated forgery information....

  43. [51]

    Experiments of Generalization Ability 3.1. Generalization to Other Unseen Datasets To validate that the accumulated forgery information en- ables our method to learn more about forgery generality, we Method DFD [10] UniFace [42] SDv15 [32] FakeA VCeleb [17] Avg. Lower Bond 0.6...

  44. [52]

    Algorithm for Sparse Uniform Replay As shown in Algorithm 1, we provide a concisely summa- rized algorithm for better comprehension in the detailed im- plementation of the proposed sparse uniform replay (SUR)

  45. [53]

    Sensitivity of replay size

    Sensitivity Evaluation Protocol 1 Protocol 2 Figure 8. Sensitivity of replay size. The shown AUCs are the average values on four datasets after training with Protocol 1 or 2. 5.1. Effect of Replay Size In Fig. 8, we examine the effect of the replay set size on model performanc...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.