Pith. sign in

REVIEW 5 major objections 5 minor 1 cited by

Pay Less Attention to Deceptive Artifacts: Robust Detection of Compressed Deepfakes on Online Social Networks

T0 review · 5 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read PLADA detects compressed deepfakes by learning to ignore JPEG block effects, reporting 76.7–77.4% mean accuracy on compressed GAN images and 80.4% on raw images.

desk verdict A plausible but undersupported SOTA claim: the method extends ODDN well, but the comparisons mix retrained and original-paper baselines and nothing has error bars. read the letter →

arxiv 2506.20548 v1 pith:AK6NNTRO submitted 2025-06-25 cs.CV cs.AIcs.LGcs.MM

Manyi Li , Renshuai Tao , Yufan Liu , Chuangchuang Tan , Haotong Qin , Bing Li , Yunchao Wei , Yao Zhao This is my paper · ORCID
classification cs.CVcs.AIcs.LGcs.MM
keywords deepfakedetectionJPEGcompressionblockeffectattentionshiftingquality-agnosticonlinesocialnetworksunpaireddatamulti-tasklearning
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

PLADA addresses a gap between deepfake detection research and real-world online images: social platforms recompress uploads with JPEG, and the resulting $8\times8$ block grid visually mimics the artifacts detectors are trained to find, degrading accuracy. The paper argues that the right response is to teach the model to pay less attention to those block effects, not to treat them as extra evidence. To do this, PLADA combines a Block Effect Eraser that shifts attention inside a pretrained transformer backbone towards deepfake traces, and an Open Data Aggregation module that clusters compressed and raw real/fake images so the real/fake boundary survives compression. On 17 GAN-generated image collections the method reports mean accuracies of 76.7% under quality-aware and 77.4% under quality-agnostic compression, roughly 4–5 points above prior methods, while retaining 80.4% on uncompressed images. If the claim holds, detectors can be built for the images people actually encounter on social platforms instead of laboratory-clean originals.

What carries the argument

Attention shifting is the load-bearing mechanism: during training, an auxiliary task predicts whether an image is compressed, and its gradient is reversed so the backbone learns to stop relying on compression fingerprints. B2E stores that guidance in per-layer guide prompt pools and injects it through two self-attention variants, Residual Guidance and Coordination Guidance, which together redirect attention away from block effects and toward deepfake artifacts at both coarse and fine scales. ODA then enforces a clustering structure in feature space, pushing real and fake centers apart within both compressed and raw groups while using the Hilbert-Schmidt Independence Criterion divergence to align a real image with its compressed counterpart. The combination is trained with a weighted sum of the real/fake loss, the compression loss, and the ODA separation loss.

What would settle it

A direct test: train PLADA on images compressed at a very high quality factor so that no block grid is visible, then compare accuracy with and without the gradient-reversed compression branch; if the branch still causes accuracy to drop, the attention shift is deleting genuine deepfake signal rather than block-effect noise. A complementary check is to train a probe classifier on PLADA's feature space and ask whether it can simultaneously ignore compression type and preserve real/fake separability; if those two objectives cannot be met together, the separability premise fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that JPEG's block effect is a deceptive artifact: it looks enough like a deepfake trace that compressed images mislead detectors, and the right countermeasure is to redirect attention away from it rather than learn it as evidence. PLADA does this with two mechanisms. The Block Effect Eraser (B2E) replaces standard self-attention in the shallow layers with two guided variants, Residual Guidance and Coordination Guidance, while a gradient-reversed compression-classification task converts the compression signal into a prompt that shifts attention toward real forgery cues. The Open Data Aggregation (ODA) module computes cluster centers for real, fake, compressed-real, and compressed-fake images and amplifies the separations, using both paired and unpaired data. The reported result is that on 17 GAN-generated datasets PLADA reaches 76.7% mean accuracy when test compression matches training and 77.4% when compression type is unknown, versus roughly 72% for the best prior method, with raw-image accuracy remaining at 80.4%.

Load-bearing premise

The account assumes that the $8\times8$ JPEG block grid leaves a feature signature separable from the traces left by deepfake generators, so steering attention away from block effects removes compression noise without erasing forgery evidence.

Editorial extensions

If this is right

  • A detector trained this way can handle unknown compression settings: quality-agnostic accuracy on 17 GAN datasets is 77.4%, about 5 points above the closest prior method.
  • Robustness to compression does not force a raw-image trade-off: PLADA stays at 80.4% mean accuracy on uncompressed images, close to baselines that ignore compression entirely.
  • The method remains effective when only 10% of training data has paired compressed copies, relaxing the need for expensive raw/compressed image pairs.
  • ODA replaces the strict paired-data machinery of HSIC-based approaches by exploiting unpaired data through cluster centers, so the training setup matches real open-world data availability.
  • Identifying block effects as a deceptive artifact gives future detectors a new target: modeling the OSN post-processing pipeline itself instead of only the generator's fingerprints.

Reading between the lines

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

  • A natural extension the paper does not test: the same attention-shifting recipe should transfer to other OSN artifacts such as resizing, sharpening, or platform overlays by swapping the auxiliary task, since the underlying assumption is that post-processing fingerprints occupy a separable feature direction.
  • If block effects are genuinely separable from deepfake traces, a lighter test-time route, projecting features away from the estimated compression direction without retraining, might recover much of PLADA's gain; the paper does not explore this cheaper option.
  • The largest average gains are reported on GAN-generated images; on some diffusion-model datasets the margin is thinner, so extending the recipe to diffusion artifacts likely needs additional guidance signals beyond the compression task.
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

5 major / 5 minor

Summary. This paper proposes PLADA, a deepfake detection framework for JPEG-compressed images, consisting of two modules: B2E (Block Effect Eraser), which redirects attention away from compression block effects via two prompt-based attention variants (RG and CG), and ODA (Open Data Aggregation), which clusters features into real/fake and raw/compressed centers and applies a distance-based loss. The authors report mean accuracies of 76.7% (quality-aware) and 77.4% (quality-agnostic) over 17 GAN datasets, competitive raw-image accuracy of 80.4%, and strong results on diffusion datasets, along with ablations on components, prompts, distance functions, and layer depths. The paper also claims that PLADA works with limited paired data and provides a public code link.

Significance. If the claimed results hold under matched conditions, PLADA would be a meaningful advance: it targets a realistic gap, namely JPEG block effects on OSNs, and shows large gains on several GAN and diffusion datasets while retaining raw-image performance. Strengths include the broad benchmark (17 GAN and 8 diffusion datasets), the public code release, and a structured ablation that separates the roles of RG, CG, and ODA. However, the central claim is explicitly comparative, and the current reporting does not yet establish that the baseline comparisons are fair or that the observed margins exceed run-to-run noise. The significance is therefore conditional on the experimental protocol being clarified and the comparisons being re-run under identical conditions.

major comments (5)
  1. [Section 4.1, Tables 1 and 3] The comparison protocol is not well-defined. The text states "Unless stated otherwise, all the above methods are retrained using our experimental settings," but the legend of Table 1 marks FF++, F3Net, MAT, SBI, QAD, and ODDN as † (results from the original paper/previous work) while Ojha, NPR, and FaT are ‡ (retrained). The same mixing appears in Tables 2–4. Because ODDN is the closest competitor, its 72.6% in Table 1 and 72.1% in Table 3 must be produced under the same 4-class ForenSynths split, the same 20% paired data with QP=50, and the same test images; otherwise the 4.1- and 5.3-point aggregate margins are not evidence of superiority. Please retrain all baselines under the identical protocol, or state for each number whether it is transferred from the original publication, and report the per-dataset comparison under matched conditions.
  2. [Section 4.2, Tables 1–6] No measure of variance or statistical significance is provided anywhere. Every table reports a single accuracy or AP value with no seeds, error bars, or significance tests. In Table 1, PLADA is below ODDN on nine of the seventeen GAN datasets (InfoGAN, CramGAN, MMD-GAN, RelGAN, SNGAN, STG-GAN, ProGAN, StyleGAN2, and CycleGAN), so the mean advantage of 76.7% over 72.6% is driven by a few large wins. Under run-to-run variability, the aggregate superiority may not be reproducible. Please report mean ± std over at least three random seeds and, if possible, a paired test over the per-dataset differences.
  3. [Section 3.4, Eq. (12)] The definition of β is circular: β appears inside L_dis in the term β Σ H, yet β is defined as σ(γ · ∂L_dis/∂Θ) ∈ (0,1), a function of L_dis itself. Moreover, ∂L_dis/∂Θ is a parameter gradient, not a scalar, so applying the sigmoid to it and requiring the result to lie in (0,1) is not well-defined. Please give a non-circular, scalar definition of β or state that it is a fixed hyperparameter; this is needed to reproduce the ODA loss.
  4. [Section 4.1, Section 3.3] The main experimental configuration is incomplete. The paper does not report the values of τ and γ in Eq. (12), the guide prompt pool size N, the guide prompt length, or the number of RG and CG layers used for the results in Tables 1–6. Figure 9 shows ablations over some of these choices but does not state which values were selected for the final model. Without this information the method cannot be reproduced from the text alone.
  5. [Section 3.3, Section 4.6] The load-bearing assumption of B2E is that the attention-shifting mechanism can suppress block-effect features while preserving deepfake traces. The only direct evidence offered is the t-SNE visualization on ProGAN and CycleGAN (Figure 10) plus downstream accuracy; this does not show which features B2E actually removes. A concrete test, e.g., comparing the frequency/spatial content of features before and after B2E or measuring block-boundary statistics, would support the causal claim that block effects are being erased rather than that the method implicitly learns a better overall classifier.
minor comments (5)
  1. [Section 3.1 and Section 4.2] The training data description says compression is set to "approximately 50% image quality," while Section 4.2 says "constant quantization parameter of 50"; these need not be the same, so please specify the exact compression setting and the codec parameters used.
  2. [Section 4.2] "ODDA trails behind" should read "ODDN trails behind"; there are also several other typographical errors, including "and ODA and" in the introduction, "the another" in Section 2, "creates noisy for deepfake detection" in Section 2.1, and "with an simple additional linear projection" in Section 4.3.
  3. [Section 4.4] The sentence "PLADA achieves the highest AP" is not supported by Table 5, which contains only accuracy columns; Table 6 reports AP only for the Ojha-test diffusion datasets, so the claim should be restricted to those results or AP columns should be added.
  4. [Figures 4 and 5] Figure 4 and Figure 5 appear to summarize many experimental conditions, but the captions do not define the axes or the color/area encoding; please make the figures self-contained.
  5. [Section 4.5.1] The ablation text gives numerical accuracies for baseline (70.14%), ODA (70.40%), RG (74.89%), RG+CG (77.86%), and full method (78.03%), but it does not state on which dataset or paired-data setting these numbers were obtained; please specify the experimental condition for the ablation study.

Circularity Check

1 steps flagged · score 2.0 of 10

No significant circularity; central claim is empirical; one self-referential loss definition noted.

  1. self definitional [Section 3.4, Eq. (12) (ODA loss definition)]
    "Ldis = Σ_{ω2∈{κ,κ}} S(φ, ϕ, ω2) + β Σ_{ω1∈{φ,ϕ}} H(C(ω1,κ), C(ω1,κ̄)) s.t. β=σ(γ·∂Ldis/∂Θ)∈(0,1)"

    The adaptive weight β is defined as a function of the gradient of Ldis with respect to Θ, while the same Ldis contains β as a coefficient multiplying the HSIC term. As written, Ldis therefore appears inside its own definition: β depends on ∂Ldis/∂Θ, and ∂Ldis/∂Θ depends on β. The paper does not state a stop-gradient, fixed-point, or detached update that would break this cycle. This is a self-referential definition of the ODA loss, though it does not by itself force the reported benchmark numbers.

full rationale

PLADA's headline claim is empirical: it reports test accuracy on ForenSynths, GANGen-Detection, Ojha-test, and DiffusionForensics, all external benchmarks. The method is not derived from a theorem, and the B2E/ODA modules are evaluated by measured accuracy. Self-citations to the authors' ODDN are present (problem definition and α=0.004), but they are not load-bearing: α is a fixed hyperparameter and the problem formulation is standard. The only structurally circular item is Eq. (12), where β is defined via the gradient of the loss that β itself weights; no detachment rule is stated. This is a genuine definitional flaw, but it is localized to a training-objective formula and does not reduce the reported gains to the input data or to the authors' prior results. Baseline-number provenance (†/‡ mixing) is a comparison-fairness concern, not circularity. Hence overall circularity is low.

Assumptions & free parameters 7 free parameters · 7 assumptions · 0 invented entities

The method relies on several domain assumptions: that JPEG compression alone simulates OSN pipelines, that attention shifting can isolate compression artifacts, and that CLIP features transfer to diffusion and GAN outputs. A set of hyperparameters (alpha, tau, gamma, prompt pool size/length, RG/CG depths) are chosen by hand or ablation without error bars; several values are not reported.

free parameters (7)
  • alpha (loss weight for L_dis) = 0.004
    Set following prior work (Tao et al., 2025), not tuned here.
  • tau (temperature in distance) = not reported
    Appears in Eq. 11; no value or selection procedure given.
  • gamma (coefficient for beta) = not reported
    Appears in Eq. 12; no value given.
  • guide prompt pool size N = small (around 4 per Fig. 9)
    Selected by ablation; not specified in text.
  • guide prompt length = 32
    Selected by ablation (Fig. 9b).
  • depth of RG layers = 16 in ablations; final uses B2E in first two layers and RG elsewhere
    Selected by ablation (Fig. 9c).
  • depth of CG layers = not specified
    Figure 9d studies it; final depth not stated.
assumptions (7)
  • domain assumption JPEG compression block effect is a key interfering artifact in OSN deepfake detection and visually resembles deepfake traces.
    Used to motivate B2E; Figure 1 and Section 1.
  • domain assumption OSN image processing can be adequately simulated by JPEG compression with quality 50 or random quality in [30,100].
    Section 3.1 and 4.1; real OSNs also resize and enhance images.
  • domain assumption Attention shifting via residual and coordination guidance can suppress compression fingerprints without destroying deepfake artifacts.
    Core premise of B2E in Section 3.3; no formal guarantee.
  • domain assumption CLIP's pretrained features generalize across GAN and diffusion model outputs.
    Backbone choice; Section 4.1.
  • domain assumption Gradient reversal of a compression classification loss injects useful guidance into prompts.
    Adversarial learning mechanism in Spark Guidance; Section 3.3.
  • domain assumption Aggregation centers and L2 distance are sufficient to align paired and unpaired compressed/raw distributions.
    ODA loss in Section 3.4; chosen over HSIC after an experiment described but not shown.
  • standard math Multi-head self-attention, backpropagation, and gradient reversal are used as standard mathematical tools.
    Sections 3.3 and 3.4 rely on these.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pay Less Attention to Deceptive Artifacts: Robust Detection of Compressed Deepfakes on Online Social Networks." pith.science (2026). https://pith.science/paper/AK6NNTRO

@misc{pith2026250620548,
  author       = {Pith},
  title        = {Pith review of: Pay Less Attention to Deceptive Artifacts: Robust Detection of Compressed Deepfakes on Online Social Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AK6NNTRO}},
  note         = {Machine review of arXiv:2506.20548}
}
read the original abstract

With the rapid advancement of deep learning, particularly through generative adversarial networks (GANs) and diffusion models (DMs), AI-generated images, or ``deepfakes", have become nearly indistinguishable from real ones. These images are widely shared across Online Social Networks (OSNs), raising concerns about their misuse. Existing deepfake detection methods overlook the ``block effects" introduced by compression in OSNs, which obscure deepfake artifacts, and primarily focus on raw images, rarely encountered in real-world scenarios. To address these challenges, we propose PLADA (Pay Less Attention to Deceptive Artifacts), a novel framework designed to tackle the lack of paired data and the ineffective use of compressed images. PLADA consists of two core modules: Block Effect Eraser (B2E), which uses a dual-stage attention mechanism to handle block effects, and Open Data Aggregation (ODA), which processes both paired and unpaired data to improve detection. Extensive experiments across 26 datasets demonstrate that PLADA achieves a remarkable balance in deepfake detection, outperforming SoTA methods in detecting deepfakes on OSNs, even with limited paired data and compression. More importantly, this work introduces the ``block effect" as a critical factor in deepfake detection, providing a robust solution for open-world scenarios. Our code is available at https://github.com/ManyiLee/PLADA.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Sliced Transport Plans

    math.OC 2025-08 unverdicted novelty 6.0 of 10

    A new sliced-Wasserstein-based discrepancy with provable semi-metric properties and a constrained Kantorovich relation, plus a generic-measure generalization of Expected Sliced plans.

Reference graph

Works this paper leans on

49 extracted references · 38 canonical work pages · cited by 1 Pith paper

  1. [1]

    write newline

    " write newline " cite write " FUNCTION editor.postfix editor num.names #1 > "( )" "( )" if FUNCTION editor.trans.postfix editor num.names #1 > "( )" "( )" if FUNCTION trans.postfix translator num.names #1 > "( )" "( )" if FUNCTION authors.editors.reflist.apa5 'field := 'dot := field num.names 'numnames := numnames 'format.num.names := format.num.names na...

  2. [2]

    ediff-i: Text-to-image diffusion models with an ensemble of expert denoisers

    Yogesh Balaji, Seungjun Nah, Xun Huang, Arash Vahdat, Jiaming Song, Qinsheng Zhang, Karsten Kreis, Miika Aittala, Timo Aila, and Samuli Laine. ediff-i: Text-to-image diffusion models with an ensemble of expert denoisers. arXiv preprint arXiv:2211.01324, 2022

    arXiv 2022

  3. [3]

    Large scale gan training for high fidelity natural image synthesis

    Andrew Brock. Large scale gan training for high fidelity natural image synthesis. arXiv preprint arXiv:1809.11096, 2018

    arXiv 2018

  4. [4]

    Towards unified defense for face forgery and spoofing attacks via dual space reconstruction learning

    Junyi Cao, Ke-Yue Zhang, Taiping Yao, Shouhong Ding, Xiaokang Yang, and Chao Ma. Towards unified defense for face forgery and spoofing attacks via dual space reconstruction learning. International Journal of Computer Vision, 132 0 (12): 0 5862--5887, 2024

    work page 2024

  5. [5]

    On the detection of synthetic images generated by diffusion models

    Riccardo Corvi, Davide Cozzolino, Giada Zingarini, Giovanni Poggi, Koki Nagano, and Luisa Verdoliva. On the detection of synthetic images generated by diffusion models. In ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1--5. IEEE, 2023. ISBN 1728163277

    work page 2023

  6. [6]

    Diffusion models beat gans on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural information processing systems, 34: 0 8780--8794, 2021

    2021

  7. [7]

    Test-time forgery detection with spatial-frequency prompt learning

    Junxian Duan, Yuang Ai, Jipeng Liu, Shenyuan Huang, Huaibo Huang, Jie Cao, and Ran He. Test-time forgery detection with spatial-frequency prompt learning. International Journal of Computer Vision, 133 0 (2): 0 672--687, 2025

    work page 2025

  8. [8]

    Fourier spectrum discrepancies in deep network generated images

    Tarik Dzanic, Karan Shah, and Freddie Witherden. Fourier spectrum discrepancies in deep network generated images. Advances in neural information processing systems, 33: 0 3022--3032, 2020

    work page 2020

Show all 49 references
  1. [9]

    Leveraging frequency analysis for deep fake image recognition

    Joel Frank, Thorsten Eisenhofer, Lea Schönherr, Asja Fischer, Dorothea Kolossa, and Thorsten Holz. Leveraging frequency analysis for deep fake image recognition. In International conference on machine learning, pages 3247--3258. PMLR, 2020. ISBN 2640-3498

    2020

  2. [10]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. Advances in neural information processing systems, 27, 2014

    2014

  3. [11]

    Language-guided hierarchical fine-grained image forgery detection and localization

    Xiao Guo, Xiaohong Liu, Iacopo Masi, and Xiaoming Liu. Language-guided hierarchical fine-grained image forgery detection and localization. International Journal of Computer Vision, 133 0 (5): 0 2670--2691, 2025

    2025

  4. [12]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. Advances in neural information processing systems, 33: 0 6840--6851, 2020

    2020

  5. [13]

    Fusing global and local features for generalized ai-synthesized image detection

    Yan Ju, Shan Jia, Lipeng Ke, Hongfei Xue, Koki Nagano, and Siwei Lyu. Fusing global and local features for generalized ai-synthesized image detection. In 2022 IEEE International Conference on Image Processing (ICIP), pages 3465--3469. IEEE, 2022. ISBN 1665496207

    2022

  6. [14]

    Alias-free generative adversarial networks

    Tero Karras, Miika Aittala, Samuli Laine, Erik Härkönen, Janne Hellsten, Jaakko Lehtinen, and Timo Aila. Alias-free generative adversarial networks. Advances in neural information processing systems, 34: 0 852--863, 2021

    2021

  7. [15]

    Quality-agnostic deepfake detection with intra-model collaborative learning

    Binh M Le and Simon S Woo. Quality-agnostic deepfake detection with intra-model collaborative learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 22378--22389, 2023

    2023

  8. [16]

    Sa 3 wt: adaptive wavelet-based transformer with self-paced auto augmentation for face forgery detection

    Yihui Li, Yifan Zhang, Hongyu Yang, Binghui Chen, and Di Huang. Sa 3 wt: adaptive wavelet-based transformer with self-paced auto augmentation for face forgery detection. International Journal of Computer Vision, 132 0 (10): 0 4417--4439, 2024

    2024

  9. [17]

    Exposing low-quality deepfake videos of social network service using spatial restored detection framework

    Ying Li, Shan Bian, Chuntao Wang, Kemal Polat, Adi Alhudhaif, and Fayadh Alenezi. Exposing low-quality deepfake videos of social network service using spatial restored detection framework. Expert Systems with Applications, 231: 0 120646, 2023. ISSN 0957-4174

    2023

  10. [18]

    Detecting generated images by real images

    Bo Liu, Fan Yang, Xiuli Bi, Bin Xiao, Weisheng Li, and Xinbo Gao. Detecting generated images by real images. In European Conference on Computer Vision, pages 95--110. Springer, 2022

    2022

  11. [19]

    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 Vision and Pattern Recognition (CVPR), pages 10770--10780, June 2024

    2024

  12. [20]

    Laa-net: Localized artifact attention network for quality-agnostic and generalizable deepfake detection

    Dat Nguyen, Nesryne Mejri, Inder Pal Singh, Polina Kuleshova, Marcella Astrid, Anis Kacem, Enjie Ghorbel, and Djamila Aouada. Laa-net: Localized artifact attention network for quality-agnostic and generalizable deepfake detection. In Proceedings of the IEEE/CVF Conference on C...

    2024

  13. [21]

    Glide: Towards photorealistic image generation and editing with text-guided diffusion models

    Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. Glide: Towards photorealistic image generation and editing with text-guided diffusion models. arXiv preprint arXiv:2112.10741, 2021

    2021 arXiv

  14. [22]

    Towards universal fake image detectors that generalize across generative models

    Utkarsh Ojha, Yuheng Li, and Yong Jae Lee. Towards universal fake image detectors that generalize across generative models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24480--24489, 2023

    2023

  15. [23]

    Pre-trained vision and language transformers are few-shot incremental learners

    Keon-Hee Park, Kyungwoo Song, and Gyeong-Moon Park. Pre-trained vision and language transformers are few-shot incremental learners. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23881--23890, 2024

    2024

  16. [24]

    Scalable diffusion models with transformers

    William Peebles and Saining Xie. Scalable diffusion models with transformers. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 4195--4205, 2023

    2023

  17. [25]

    Thinking in frequency: Face forgery detection by mining frequency-aware clues

    Yuyang Qian, Guojun Yin, Lu Sheng, Zixuan Chen, and Jing Shao. Thinking in frequency: Face forgery detection by mining frequency-aware clues. In European conference on computer vision, pages 86--103. Springer, 2020

    2020

  18. [26]

    Df-rap: A robust adversarial perturbation for defending against deepfakes in real-world social network scenarios

    Zuomin Qu, Zuping Xi, Wei Lu, Xiangyang Luo, Qian Wang, and Bin Li. Df-rap: A robust adversarial perturbation for defending against deepfakes in real-world social network scenarios. IEEE Transactions on Information Forensics and Security, 2024. ISSN 1556-6013

    2024

  19. [27]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, and Jack Clark. Learning transferable visual models from natural language supervision. In International conference on machine learning, pages...

    2021

  20. [28]

    Hierarchical text-conditional image generation with clip latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125, 1 0 (2): 0 3, 2022

    2022 arXiv

  21. [29]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684--10695, 2022

    2022

  22. [30]

    Faceforensics++: Learning to detect manipulated facial images

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

    2019

  23. [31]

    Stylegan-t: Unlocking the power of gans for fast large-scale text-to-image synthesis

    Axel Sauer, Tero Karras, Samuli Laine, Andreas Geiger, and Timo Aila. Stylegan-t: Unlocking the power of gans for fast large-scale text-to-image synthesis. In International conference on machine learning, pages 30105--30118. PMLR, 2023. ISBN 2640-3498

    2023

  24. [32]

    Robust sequential deepfake detection

    Rui Shao, Tianxing Wu, and Ziwei Liu. Robust sequential deepfake detection. International Journal of Computer Vision, pages 1--18, 2025

    2025

  25. [33]

    Interpreting the latent space of gans for semantic face editing

    Yujun Shen, Jinjin Gu, Xiaoou Tang, and Bolei Zhou. Interpreting the latent space of gans for semantic face editing. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9243--9252, 2020

    2020

  26. [34]

    Detecting deepfakes with self-blended images

    Kaede Shiohara and Toshihiko Yamasaki. Detecting deepfakes with self-blended images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18720--18729, 2022

    2022

  27. [35]

    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. International Journal of Computer Vision, 133 0 (3): 0 1067--1084, 2025

    2025

  28. [36]

    Processing-aware privacy-preserving photo sharing over online social networks

    Weiwei Sun, Jiantao Zhou, Ran Lyu, and Shuyuan Zhu. Processing-aware privacy-preserving photo sharing over online social networks. In Proceedings of the 24th ACM international conference on Multimedia, pages 581--585, 2016

    2016

  29. [37]

    Gangen-detection: A dataset generated by gans for generalizable deepfake detection

    Chuangchuang Tan and Renshuai Tao. Gangen-detection: A dataset generated by gans for generalizable deepfake detection. https://github.com/chuangchuangtan/GANGen-Detection, 2024

    2024

  30. [38]

    Rethinking the up-sampling operations 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 operations in cnn-based generative network for generalizable deepfake detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pag...

    2024

  31. [39]

    Mind the interference: Retaining pre-trained knowledge in parameter efficient continual learning of vision-language models

    Longxiang Tang, Zhuotao Tian, Kai Li, Chunming He, Hantao Zhou, Hengshuang Zhao, Xiu Li, and Jiaya Jia. Mind the interference: Retaining pre-trained knowledge in parameter efficient continual learning of vision-language models. In European Conference on Computer Vision, pages ...

    2025

  32. [40]

    Oddn: Addressing unpaired data challenges in open-world deepfake detection on online social networks

    Renshuai Tao, Manyi Le, Chuangchuang Tan, Huan Liu, Haotong Qin, and Yao Zhao. Oddn: Addressing unpaired data challenges in open-world deepfake detection on online social networks. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 799--807, 2025

    2025

  33. [41]

    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. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8695--8704, 2020

    2020

  34. [42]

    Watcher: Wavelet-guided texture-content hierarchical relation learning for deepfake detection

    Yuan Wang, Chen Chen, Ning Zhang, and Xiyuan Hu. Watcher: Wavelet-guided texture-content hierarchical relation learning for deepfake detection. International Journal of Computer Vision, 132 0 (10): 0 4746--4767, 2024

    2024

  35. [43]

    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. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 22445--22455, 2023

    2023

  36. [44]

    Fd-gan: Generalizable and robust forgery detection via generative adversarial networks

    Nanqing Xu, Weiwei Feng, Tianzhu Zhang, and Yongdong Zhang. Fd-gan: Generalizable and robust forgery detection via generative adversarial networks. International Journal of Computer Vision, 132 0 (12): 0 5801--5819, 2024 a

    2024

  37. [45]

    Learning pairwise interaction for generalizable deepfake detection

    Ying Xu, Kiran Raja, Luisa Verdoliva, and Marius Pedersen. Learning pairwise interaction for generalizable deepfake detection. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, pages 672--682, 2023

    2023

  38. [46]

    Learning spatiotemporal inconsistency via thumbnail layout for face deepfake detection

    Yuting Xu, Jian Liang, Lijun Sheng, and Xiao-Yu Zhang. Learning spatiotemporal inconsistency via thumbnail layout for face deepfake detection. International Journal of Computer Vision, 132 0 (12): 0 5663--5680, 2024 b

    2024

  39. [47]

    Mining generalized multi-timescale inconsistency for detecting deepfake videos

    Yang Yu, Rongrong Ni, Siyuan Yang, Yu Ni, Yao Zhao, and Alex C Kot. Mining generalized multi-timescale inconsistency for detecting deepfake videos. International Journal of Computer Vision, pages 1--17, 2024

    2024

  40. [48]

    Multi-attentional deepfake detection

    Hanqing Zhao, Wenbo Zhou, Dongdong Chen, Tianyi Wei, Weiming Zhang, and Nenghai Yu. Multi-attentional deepfake detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2185--2194, 2021

    2021

  41. [49]

    write newline

    " write newline "" before.all 'output.state := FUNCTION output.doi doi empty skip "doi:" doi * "" * output if FUNCTION format.archive archivePrefix empty "" archivePrefix ":" * if FUNCTION format.primaryClass primaryClass empty "" " [" primaryClass * "] " * if FUNCTION format....

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.