Pith. sign in

REVIEW 3 major objections 5 minor 40 references

Bridging the Gap: A Framework for Real-World Video Deepfake Detection via Social Network Compression Emulation

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Detectors fine-tuned on locally emulated social-network compression match those trained on real platform uploads.

desk verdict A useful, practical emulator that mostly delivers, but the missing generic-compression baseline leaves the platform-specific claim under-supported. read the letter →

arxiv 2508.08765 v2 pith:Q47UGQ3C submitted 2025-08-12 cs.CV cs.AI

Andrea Montibeller , Dasara Shullani , Daniele Baracchi , Alessandro Piva , Giulia Boato This is my paper · ORCID
classification cs.CVcs.AI
keywords videodeepfakedetectionsocialnetworkcompressionemulationconstantratefactorestimationH.264re-encodingfine-tuningforrobustnessmultimediaforensics
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

The paper argues that the gap between lab-trained deepfake detectors and real-world performance is mostly caused by the aggressive, proprietary compression that social networks apply to uploaded video, and that this compression can be reproduced locally without API access. To show this, it builds the Social Network Video Sharing Emulator (SNVSE), which estimates, from fewer than fifty uploaded videos per resolution, the output resolution and average constant-rate factor a platform applies, then re-encodes any local video with H.264 to match. On a standard face-manipulation benchmark, detectors fine-tuned on these emulated videos perform comparably to detectors fine-tuned on videos actually shared on Facebook, YouTube, and BlueSky, for both 2021-era and 2025-era platform pipelines. If correct, the result makes realistic deepfake training data cheap and API-free, and gives a practical way to keep detectors current as platforms change their encoders.

What carries the argument

The central machinery is the two-module Social Network Video Sharing Emulator (SNVSE). A Parameters Estimation Module takes pairs of original and platform-shared videos and, for each pair, finds the smallest integer CRF in 21–50 whose H.264 re-encode at the shared video's resolution has bitrate at most that of the shared video, storing triplets (input resolution, output resolution, estimated CRF). An Encoding Emulation Module then maps any new video to the nearest stored input resolution, takes the output resolution from that entry, and re-encodes with the average CRF across entries sharing that output resolution. This reduces platform-specific transcoding to one resolution mapping plus one averaged CRF per output resolution, which is what makes large-scale local emulation tractable.

What would settle it

Take a set of videos covering varied content, share them on platforms whose encoders use content-adaptive rate control or VP9/AV1, estimate CRFs with the paper's bitrate-matching rule, and compare detector true-negative rates after fine-tuning on SNVSE emulations versus on actual uploads; if the emulation fine-tuning falls measurably short on those platforms, the averaged-CRF model is the point of failure. The paper's own NeuralTextures/YouTube discrepancy in Fig. 2b is a candidate early signal for such a failure.

Watch

Extended reading notes

Core claim

The paper's central claim is that platform-specific social-network compression can be reduced to two scalar parameters—an output resolution and an average constant rate factor (CRF)—estimated from a small set of original/shared video pairs, and that re-encoding with those parameters reproduces the degradation that matters for deepfake detectors. The authors evaluate this on a standard face-manipulation dataset with real and fake videos shared on Facebook and YouTube in 2021 and on BlueSky, Facebook, and YouTube in 2025. They report that detectors fine-tuned on emulated videos recover most of the accuracy lost when testing on actual shared videos, and in most configurations match detectors fine-tuned on the true shared media. The one reported shortfall is the NeuralTextures manipulation on YouTube, where emulated fine-tuning improves true-negative rate less than shared fine-tuning does; the authors attribute this to fixed-threshold evaluation rather than to a failure of the emulator.

Load-bearing premise

The load-bearing premise is that re-encoding a video once with H.264 at a fixed quality setting and a changed resolution captures the visual degradation a platform's own transcoding produces, well enough that fine-tuning on those re-encoded videos transfers to actual platform uploads.

Editorial extensions

If this is right

  • Fine-tuning on SNVSE-emulated videos restores most of the true-negative rate lost when a detector trained on unshared video meets Facebook- or YouTube-compressed fakes.
  • The framework needs fewer than fifty shared videos per resolution; the bootstrap ablation indicates at least thirty per resolution for stable CRF estimates.
  • Emulation transfers to current platform pipelines (2025 Facebook, YouTube, and BlueSky) as well as to the original 2021-era uploads.
  • Because emulation runs locally, any existing video dataset can be processed at scale without API quotas, sharing limits, or per-video costs.
  • Videos at resolutions absent from the database are handled by nearest-neighbor resolution matching plus averaged CRF, so coverage is not limited to the exact resolutions observed.

Reading between the lines

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

  • If a single averaged CRF per resolution is enough, the same parameter-estimation trick should extend to other codecs and to image sharing; platforms using VP9 or AV1, or content-adaptive rate control, are the natural stress test for the averaging assumption.
  • The framework implicitly predicts that a detector fine-tuned on emulated data from several platforms will generalize across platforms; that cross-platform generalization is not directly tested and would be a cheap extension.
  • One could use the emulator to quantify how much of the lab-to-real gap is compression alone versus other platform effects, such as watermarking, container changes, or audio re-encoding, by ablating which emulated factors are applied.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes SNVSE, a framework that emulates social-network video compression for deepfake detection. From a small set of paired original/shared videos, it estimates a per-video CRF by matching the bitrate of an H.264 re-encode to the shared video (Eq. 1), stores input/output resolution and CRF triplets (Eq. 2), and for new videos selects an output resolution by nearest match and a CRF by averaging over the database for that resolution (Eqs. 3-4). The emulated videos are then used to fine-tune four CNN detectors on FF++ data, and the paper reports TPR/TNR on test videos shared via Facebook, YouTube, and BlueSky in 2021 and 2025. The main claims are that the emulated videos closely match real platform degradation and that fine-tuning on emulated videos gives performance comparable to fine-tuning on actually shared videos. A bootstrap ablation suggests that at least 30 shared videos per resolution are needed for stable CRF estimates.

Significance. If the central claim is established, the framework is practically valuable: it would allow scalable generation of platform-like compressed training data without API access, and the code release is a concrete contribution. The evaluation design is partially sound in that CRF parameters are estimated from validation videos and tested on a held-out test set, so the setup is not circular. The extension to 2025 platform versions and BlueSky is a useful generalization check. However, the evidence currently falls short of the fidelity claim: there is no direct comparison of emulated versus shared video quality, and no baseline against generic H.264 re-encoding, which is necessary to show that the estimated platform parameters, rather than compression in general, drive the observed recovery.

major comments (3)
  1. [Sec. 4.1, Fig. 2] The experimental protocol compares NS, S, and EMU fine-tuning but never compares against generic H.264 re-encoding at matched resolutions. Since an emulated video is exactly an H.264 re-encode with a resolution change and a CRF taken from [21,50], any compression pipeline could introduce artifacts that a detector latches onto. If a fixed-CRF or random-CRF re-encode produces the same TNR recovery, then the platform-specific parameter estimation in Eqs. (1)-(4) is not load-bearing. Please add this negative control and report whether the two-parameter model outperforms it.
  2. [Abstract; Sec. 4.1] The abstract's claim that emulated data 'closely matches the degradation patterns of real uploads' is not directly measured anywhere in the paper. Section 4.1 reports only downstream detector TPR and TNR; there is no PSNR/SSIM comparison, no bitrate or artifact-distribution analysis, and no visual or forensic-feature comparison between emulated and shared videos. Task-level equivalence is a necessary condition, but it does not by itself establish fidelity of the degradation patterns. Please add a direct fidelity check or soften the claim.
  3. [Fig. 2b, Sec. 4.1] The 'comparable performance' claim rests on point estimates without error bars or significance tests, and one cell directly contradicts it: for NeuralTextures on YouTube, EMU TNR is consistently below S TNR across all four detectors. The paper attributes this to sub-optimal thresholding but provides no verification. Please provide confidence intervals or error bars, and either verify the thresholding explanation or restrict the claim to the cells where it holds.
minor comments (5)
  1. [Eq. (1)] If no CRF in [21,50] satisfies the bitrate constraint, the minimization in Eq. (1) is over an empty set; please define the fallback behavior explicitly.
  2. [Sec. 4.2, Fig. 3] The text says the 2025 results are 'consistent' with Section 4.1, but the Facebook and YouTube cells are based on only 50 validation videos per class and show visible gaps; please qualify this statement and note the reduced statistical power.
  3. [Sec. 3.1 and Sec. 5] Frame rate and pixel format are said in Section 5 to be future work, but this limitation should be stated in Section 3.1 where the encoding parameters are defined, since the emulator currently covers only a subset of encoding decisions.
  4. [References [16, 17, 19]] The references for DenseNet, InceptionNet, and XceptionNet are application papers rather than the original architecture papers; please cite the original sources.
  5. [Sec. 4] The fine-tuning procedure is described only as following [23]; please report the learning rate, number of epochs, batch size, and any data augmentation used, even if they match the prior work, to make the experiments reproducible without consulting the previous paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CRF/resolution parameters are fitted to a validation set and evaluated on held-out shared videos via detector TPR/TNR, an external benchmark.

full rationale

The paper's derivation chain is not circular. The only fitted quantities are the CRF values c_i (Eq. 1), estimated per validation video by matching the bitrate of a re-encode to the corresponding shared video, and the per-resolution average c* (Eq. 4). These parameters are estimated from the validation split of FF++ shared videos and are then applied to re-encode the original (non-shared) validation videos to produce EMU data. The central claim—that fine-tuning on EMU data yields detector performance comparable to fine-tuning on actual shared videos—is assessed on the held-out FF++ test set of real shared videos, using TPR/TNR of four CNN detectors. The evaluated labels and videos are not used in estimating c_i or c*, so the result is not forced by construction. The self-citations ([15],[22],[23]) provide the shared-video dataset and evaluation protocol, which are external, reproducible data rather than unverified theorems or fitted values; they do not import the conclusion. The absence of a generic-H.264 negative control and the lower EMU TNR for NeuralTextures are experimental-validity concerns about whether the two-parameter emulator is the active ingredient, but they do not make the reported comparison equivalent to its inputs. The Sec. 5 admission that frame rate and pixel format are not estimated limits scope but does not create circularity.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The method's parameters are all fitted to the observed shared-video bitrates; the only check on their validity is downstream detector performance, which is itself noisy. The central claim therefore depends on the assumption that compression artifacts are captured by a single CRF per resolution.

free parameters (5)
  • CRF search range (c_min, c_max) = 21 to 50
    Hand-chosen bounds for the bitrate-matching sweep in Eq. (1); the choice affects which c_hat values can be found.
  • Per-video estimated CRF c_hat_i = determined per video by Eq. (1)
    Fitted to each shared video by requiring re-encoded bitrate <= observed shared bitrate; this is the core data-fitting step.
  • Mean CRF c*_j per output resolution = average of c_hat_l over l with same output resolution
    Defined in Eq. (4); this aggregated value is applied to all new videos at that resolution, assuming content-independence.
  • Resolution matching rule = nearest input resolution via Euclidean distance (Eq. 3)
    Design choice for mapping unseen input resolutions to stored output resolutions; no validation of optimality.
  • Minimum videos per resolution = 30
    Derived from bootstrap ablation on 1280x720 FF++ validation videos (Sec. 4.3); generalizability to other resolutions/platforms is assumed.
assumptions (4)
  • domain assumption Social network video processing is reducible to a single H.264 re-encode with a CRF and output resolution
    The emulator models platform transcoding via Eqs. (1)-(4); no evidence that platforms use single-pass CRF H.264.
  • domain assumption Bitrate matching via CRF sweep produces artifact-equivalent compression
    Eq. (1) equates compression behavior with bitrate; different encoders can yield similar bitrate but different artifacts.
  • domain assumption Average CRF per output resolution transfers to unseen videos
    Eq. (4) assumes content variability in CRF is captured by the mean; the NT/YouTube gap suggests this can fail.
  • domain assumption FF++ shared-video protocol from Marcon et al. is a valid benchmark
    All experiments reuse this protocol; its shared videos were uploaded in 2021 and may not reflect current pipelines.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Bridging the Gap: A Framework for Real-World Video Deepfake Detection via Social Network Compression Emulation." pith.science (2026). https://pith.science/paper/Q47UGQ3C

@misc{pith2026250808765,
  author       = {Pith},
  title        = {Pith review of: Bridging the Gap: A Framework for Real-World Video Deepfake Detection via Social Network Compression Emulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q47UGQ3C}},
  note         = {Machine review of arXiv:2508.08765}
}
read the original abstract

The growing presence of AI-generated videos on social networks poses new challenges for deepfake detection, as detectors trained under controlled conditions often fail to generalize to real-world scenarios. A key factor behind this gap is the aggressive, proprietary compression applied by platforms like YouTube and Facebook, which launder low-level forensic cues. However, replicating these transformations at scale is difficult due to API limitations and data-sharing constraints. For these reasons, we propose a first framework that emulates the video sharing pipelines of social networks by estimating compression and resizing parameters from a small set of uploaded videos. These parameters enable a local emulator capable of reproducing platform-specific artifacts on large datasets without direct API access. Experiments on FaceForensics++ videos shared via social networks demonstrate that our emulated data closely matches the degradation patterns of real uploads. Furthermore, detectors fine-tuned on emulated videos achieve comparable performance to those trained on actual shared media. Our approach offers a scalable and practical solution for bridging the gap between lab-based training and real-world deployment of deepfake detectors, particularly in the underexplored domain of compressed video content.

Figures

Figures reproduced from arXiv: 2508.08765 by the authors.

Figure 1
Figure 1. Overview of the Social Network Video Sharing Em [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. TPR and TNR obtained by four deepfake detectors on videos shared on Facebook and YouTube. Models were trained [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. TPR and TNR obtained by detectors evaluated on videos shared, during 2025, on BlueSky, YouTube, and Facebook. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Stability analysis of CRF parameter estimation across bootstrap validation, using shared videos from FF++ [ [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 33 canonical work pages

  1. [1]

    High-Resolution Image Synthesis with Latent Diffusion Models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High-Resolution Image Synthesis with Latent Diffusion Models, ” Apr. 2022. arXiv:2112.10752

    arXiv 2022

  2. [2]

    Sdxl: Improving latent diffusion models for high-resolution image synthesis,

    D. Podell, Z. English, K. Lacey, A. Blattmann, T. Dockhorn, J. Müller, J. Penna, and R. Rombach, “Sdxl: Improving latent diffusion models for high-resolution image synthesis, ”arXiv preprint arXiv:2307.01952, 2023

    arXiv 2023

  3. [3]

    B. F. Labs, “Flux. ” https://github.com/black-forest-labs/flux, 2024

    work page 2024

  4. [4]

    Deepfloyd if: A powerful open source text-to-image model,

    D. Research, “Deepfloyd if: A powerful open source text-to-image model, ” 2023. https://github.com/deep-floyd/IF

    work page 2023

  5. [5]

    Kandinsky 2.2: Open-source text-to-image and image-to- image model,

    P. Kuznetsovet al., “Kandinsky 2.2: Open-source text-to-image and image-to- image model, ” 2023. https://github.com/ai-forever/Kandinsky-2

    work page 2023

  6. [6]

    Taming transformers for high-resolution image synthesis,

    P. Esser, R. Rombach, and B. Ommer, “Taming transformers for high-resolution image synthesis, ”CVPR, 2021

    work page 2021

  7. [7]

    Animatediff: Animate your personalized text-to-image diffusion models without specific tuning,

    Y. Guo, C. Yang, A. Rao, Z. Liang, Y. Wang, Y. Qiao, M. Agrawala, D. Lin, and B. Dai, “Animatediff: Animate your personalized text-to-image diffusion models without specific tuning, ” 2023. https://github.com/guoyww/AnimateDiff

    work page 2023

  8. [8]

    Audioldm: text-to-audio generation with latent diffusion models,

    H. Liu, Z. Chen, Y. Yuan, X. Mei, X. Liu, D. Mandic, W. Wang, and M. D. Plumbley, “Audioldm: text-to-audio generation with latent diffusion models, ” inProceedings of the 40th International Conference on Machine Learning, ICML’23, JMLR.org, 2023

    work page 2023

Show all 40 references
  1. [9]

    Advance fake video detection via vision transformers,

    J. Battocchio, S. Dell’Anna, A. Montibeller, and G. Boato, “Advance fake video detection via vision transformers, ” inProceedings of the ACM Workshop on Infor- mation Hiding and Multimedia Security, pp. 1–11, 2025

    2025

  2. [10]

    Face2face: Real-time face capture and reenactment of rgb videos,

    J. Thies, M. Zollhofer, M. Stamminger, C. Theobalt, and M. Nießner, “Face2face: Real-time face capture and reenactment of rgb videos, ” inProceedings of the IEEE conference on computer vision and pattern recognition, pp. 2387–2395, 2016

    2016

  3. [11]

    Asvspoof 2019: Future horizons in spoofed and fake audio detection,

    M. Todisco, X. Wang, V. Vestman, M. Sahidullah, H. Delgado, A. Nautsch, J. Yam- agishi, N. Evans, T. Kinnunen, and K. A. Lee, “Asvspoof 2019: Future horizons in spoofed and fake audio detection, ”arXiv preprint arXiv:1904.05441, 2019

    2019 arXiv

  4. [12]

    Beyond the brush: Fully-automated crafting of realistic inpainted images,

    G. Bertazzini, C. Albisani, D. Baracchi, D. Shullani, and A. Piva, “Beyond the brush: Fully-automated crafting of realistic inpainted images, ” in2024 IEEE International Workshop on Information Forensics and Security (WIFS), pp. 1–6, IEEE, 2024

    2024

  5. [13]

    Deepfake media forensics: State of the art and challenges ahead,

    I. Amerini, M. Barni, S. Battiato, P. Bestagini, G. Boato, T. S. Bonaventura, V. Bruni, R. Caldelli, F. De Natale, R. De Nicola,et al., “Deepfake media forensics: State of the art and challenges ahead, ”arXiv preprint arXiv:2408.00388, 2024

    2024 arXiv

  6. [14]

    Deepfake media forensics: Status and future challenges,

    I. Amerini, M. Barni, S. Battiato, P. Bestagini, G. Boato, V. Bruni, R. Caldelli, F. De Natale, R. De Nicola, L. Guarnera,et al., “Deepfake media forensics: Status and future challenges, ”Journal of Imaging, vol. 11, no. 3, p. 73, 2025

    2025

  7. [15]

    Truefake: A real world case dataset of last generation fake images also shared on social networks,

    S. Dell’Anna, A. Montibeller, and G. Boato, “Truefake: A real world case dataset of last generation fake images also shared on social networks, ”arXiv preprint arXiv:2504.20658, 2025

    2025

  8. [16]

    Resnet 50,

    B. Koonce and B. Koonce, “Resnet 50, ”Convolutional Neural Networks with Swift for TensorFlow: Image Recognition and Dataset Categorization, pp. 63–72, 2021

    2021

  9. [17]

    Classification of paediatric pneumonia using modified densenet- 121 deep-learning model,

    T. Arulananth, S. W. Prakash, R. K. Ayyasamy, V. Kavitha, P. Kuppusamy, and P. Chinnasamy, “Classification of paediatric pneumonia using modified densenet- 121 deep-learning model, ”IEEE Access, 2024

    2024

  10. [18]

    Efficientnet,

    B. Koonce, “Efficientnet, ” inConvolutional neural networks with swift for Tensor- flow: image recognition and dataset categorization, pp. 109–123, Springer, 2021

    2021

  11. [19]

    Deep learning-based classification for melanoma detection using xceptionnet,

    X. Lu and Y. Firoozeh Abolhasani Zadeh, “Deep learning-based classification for melanoma detection using xceptionnet, ”Journal of Healthcare Engineering, vol. 2022, no. 1, p. 2196096, 2022. DFF ’25, October 27–28, 2025, Dublin, Ireland Andrea Montibeller, Dasara Shullani, Dani...

    2022

  12. [20]

    Raising the Bar of AI-generated Image Detection with CLIP,

    D. Cozzolino, G. Poggi, R. Corvi, M. Nießner, and L. Verdoliva, “Raising the Bar of AI-generated Image Detection with CLIP, ” in2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp. 4356–4366, June 2024. ISSN: 2160-7516

    2024

  13. [21]

    Low-rank adaptation method for wav2vec2-based fake audio detection,

    C. Wang, J. Yi, X. Zhang, J. Tao, L. Xu, and R. Fu, “Low-rank adaptation method for wav2vec2-based fake audio detection, ”arXiv preprint arXiv:2306.05617, 2023

    2023 arXiv

  14. [22]

    TrueFace: a Dataset for the Detection of Synthetic Face Images from Social Networks,

    G. Boato, C. Pasquini, A. L. Stefani, S. Verde, and D. Miorandi, “TrueFace: a Dataset for the Detection of Synthetic Face Images from Social Networks, ” in 2022 IEEE International Joint Conference on Biometrics (IJCB), pp. 1–7, Oct. 2022. ISSN: 2474-9699

    2022

  15. [23]

    Detection of manipulated face videos over social networks: A large-scale study,

    F. Marcon, C. Pasquini, and G. Boato, “Detection of manipulated face videos over social networks: A large-scale study, ”Journal of Imaging, vol. 7, no. 10, p. 193, 2021

    2021

  16. [24]

    Detection of gan- generated fake images over social networks,

    F. Marra, D. Gragnaniello, D. Cozzolino, and L. Verdoliva, “Detection of gan- generated fake images over social networks, ” in2018 IEEE Conference on Multi- media Information Processing and Retrieval (MIPR), pp. 384–389, 2018

    2018

  17. [25]

    Media forensics on social media platforms: a survey,

    C. Pasquini, I. Amerini, and G. Boato, “Media forensics on social media platforms: a survey, ”EURASIP Journal on Information Security, vol. 2021, no. 1, p. 4, 2021

    2021

  18. [26]

    Media forensics and deepfakes: An overview,

    L. Verdoliva, “Media forensics and deepfakes: An overview, ”IEEE Journal on Selected Topics in Signal Processing, vol. 14, no. 5, p. 910 – 932, 2020

    2020

  19. [27]

    Reliable out-of-distribution recognition of synthetic images,

    A. Maier and C. Riess, “Reliable out-of-distribution recognition of synthetic images, ”Journal of Imaging, vol. 10, no. 5, p. 110, 2024

    2024

  20. [28]

    Characteristics and prevalence of fake social media profiles with ai-generated faces,

    K.-C. Yang, D. Singh, and F. Menczer, “Characteristics and prevalence of fake social media profiles with ai-generated faces, ”arXiv Preprint arXiv:2401.02627, 2024

    2024 arXiv

  21. [29]

    Ai-generated faces in the real world: a large-scale case study of twitter profile images,

    J. Ricker, D. Assenmacher, T. Holz, A. Fischer, and E. Quiring, “Ai-generated faces in the real world: a large-scale case study of twitter profile images, ” in Proceedings of the 27th International Symposium on Research in Attacks, Intrusions and Defenses, pp. 513–530, 2024

    2024

  22. [30]

    A gan-based model of deepfake detection in social media,

    M. Kumar, H. K. Sharma,et al., “A gan-based model of deepfake detection in social media, ”Procedia Computer Science, vol. 218, pp. 2153–2162, 2023

    2023

  23. [31]

    A style-based generator architecture for genera- tive adversarial networks,

    T. Karras, S. Laine, and T. Aila, “A style-based generator architecture for genera- tive adversarial networks, ” in2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 4396–4405, 2019

    2019

  24. [32]

    Catastrophic forgetting in connectionist networks,

    R. M. French, “Catastrophic forgetting in connectionist networks, ”Trends in cognitive sciences, vol. 3, no. 4, pp. 128–135, 1999

    1999

  25. [33]

    Detecting compressed deepfake videos in social networks using frame-temporality two-stream convolutional network,

    J. Hu, X. Liao, W. Wang, and Z. Qin, “Detecting compressed deepfake videos in social networks using frame-temporality two-stream convolutional network, ” IEEE Transactions on Circuits and Systems for Video Technology, vol. 32, no. 3, pp. 1089–1102, 2021

    2021

  26. [34]

    Face- Forensics++: Learning to Detect Manipulated Facial Images,

    A. Rössler, D. Cozzolino, L. Verdoliva, C. Riess, J. Thies, and M. Niessner, “Face- Forensics++: Learning to Detect Manipulated Facial Images, ” in2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 1–11, Oct. 2019. ISSN: 2380-7504

    2019

  27. [35]

    faceswap

    deepfakes, “faceswap. ” GitHub repository, 2018. Accessed: 2018-10-29

    2018

  28. [36]

    Faceswap

    Kowalski, Marek, “Faceswap. ” GitHub repository, 2018. Accessed: 2018-10-29

    2018

  29. [37]

    Faceshifter: Towards high fidelity and occlusion aware face swapping,

    L. Li, J. Bao, H. Yang, D. Chen, and F. Wen, “Faceshifter: Towards high fidelity and occlusion aware face swapping, ”arXiv preprint arXiv:1912.13457, 2019

    1912 arXiv

  30. [38]

    Deferred neural rendering: Image syn- thesis using neural textures,

    J. Thies, M. Zollhöfer, and M. Nießner, “Deferred neural rendering: Image syn- thesis using neural textures, ”Acm Transactions on Graphics (TOG), vol. 38, no. 4, pp. 1–12, 2019

    2019

  31. [39]

    Rethinking the inception architecture for computer vision,

    C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna, “Rethinking the inception architecture for computer vision, ” inProceedings of the IEEE conference on computer vision and pattern recognition, pp. 2818–2826, 2016

    2016

  32. [40]

    Bootstrap methods: another look at the jackknife,

    B. Efron, “Bootstrap methods: another look at the jackknife, ” inBreakthroughs in statistics: Methodology and distribution, pp. 569–593, Springer, 1992

    1992

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.