Pith. sign in

REVIEW 3 major objections 7 minor 2 cited by

Learning Real Facial Concepts for Independent Deepfake Detection

T0 review · 3 major / 7 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 RealID, trained only on FaceForensics++, achieves 91.06 average AUC across five unseen deepfake datasets by learning diverse real-face prototypes and deciding real versus fake independently.

desk verdict RealID is a plausible new architecture for cross-dataset deepfake detection, but the headline 1.74% average AUC gain over SBI is compromised by test-set hyperparameter tuning. read the letter →

arxiv 2505.04460 v1 pith:2F2WX52Q submitted 2025-05-07 cs.CV

Ming-Hui Liu , Harry Cheng , Tianyi Wang , Xin Luo , Xin-Shun Xu This is my paper · ORCID
classification cs.CV
keywords deepfakedetectioncross-datasetgeneralizationrealfacialconceptsprototypelearningmemorynetworksindependentdual-decisionclassifierFaceForensics++AUC
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

RealID claims that deepfake detectors fail on unseen datasets because they learn a single, narrow notion of real faces and lean too heavily on forgery artifacts, which leads them to label genuine faces as fake. To fix this, it stores several prototypes of real faces in a memory bank and trains a classifier that scores the real and fake hypotheses independently instead of as two sides of one coin. Trained only on FaceForensics++, RealID raises the average cross-dataset AUC across Celeb-DF, DFD, DFDC, DFDCp, and UADFV to 91.06, beating the strongest prior method by 1.74 average AUC. The implied payoff is a detector whose decisions transfer to forgery types and image conditions it has never seen.

What carries the argument

The load-bearing object is the pair (RealC2, IDC). RealC2 is a memory bank of $K$ real-face prototypes, updated with a softmax-weighted combination of the real features assigned to each prototype, plus two losses: $\mathcal{L}_{\text{Distinction}}$ pulls each real feature toward its nearest prototype and the mean real prototype while pushing it away from fake prototypes, and $\mathcal{L}_{\text{Diversity}}$ keeps the second-nearest prototype at least a margin away. IDC is a classifier with four outputs and regularization $R$: for a real sample the term pushes the fake probability toward the auxiliary output $\hat f_j^3$, and for a fake sample it pushes the real probability toward $\hat f_j^2$, making the two decisions independent. The total loss is $\mathcal{L}_{\text{CE}}+\lambda_1\mathcal{L}_{\text{Diversity}}+\lambda_2\mathcal{L}_{\text{Distinction}}+\lambda_3 R$.

What would settle it

Run RealID and SBI on FF++ exactly as reported, then evaluate both on a curated deepfake dataset that was not touched during development and not derived from any dataset used for tuning or ablations; if the average AUC gap over SBI is smaller than 1.74 or reverses, the reported advantage was tuned to the five test sets rather than a general property.

Watch

Extended reading notes

Core claim

The paper's central claim is that generalization in deepfake detection improves when the model builds a comprehensive concept of real faces and applies it independently of forgery cues. The Real Concept Capture module (RealC2) maintains $K$ real prototypes in a Multi-Real Memory, updating each prototype from the real samples assigned to it and enforcing both intra-subclass tightness (Prototype Distinction Loss) and spread between prototypes (Prototype Diversity Loss). The Independent Dual-Decision Classifier (IDC) replaces the binary output with a four-dimensional softmax, $\left(\hat f_j^0, \hat f_j^1, \hat f_j^2, \hat f_j^3\right)$, whose first two entries are still real and fake probabilities but whose auxiliary entries absorb error through a regularization term, so a rise in the fake score no longer forces a fall in the real score. With both modules, the model can reject a sample because it lacks the learned real-face concept rather than because some local artifact resembles a forgery, and the reported cross-dataset results support that mechanism.

Load-bearing premise

The head-to-head claim stands on the assumption that the three loss weights $\lambda_1,\lambda_2,\lambda_3$ were fixed without consulting the five test sets; the paper grid-searches them and reports Figure 4 varying each weight while evaluating on Celeb-DF and DFDC, two of the five evaluation datasets, so the 1.74 average AUC margin is an in-sample number unless a separate validation split was used.

Editorial extensions

If this is right

  • If the two-module design is responsible for the gain, then any detector that learns a multi-prototype model of the real class and decouples the real/fake decisions should transfer better, not just this architecture.
  • The 1.74 average AUC gain over SBI on five datasets suggests the main remaining generalization bottleneck is the real class distribution, not richer forgery augmentation.
  • Backbone-agnostic improvements (Xception, ViT-L, ViT-B all gain from RealID) mean the module can be grafted onto stronger feature extractors as they appear.
  • Heatmaps that spread attention across lips, nose, and eyes, rather than local traces, imply the detector is using a holistic real-face criterion; video-level aggregation may amplify this advantage.

Reading between the lines

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

  • The paper does not test this, but a natural extension is applying RealID to video-level detection by aggregating frame-level independent real/fake scores, which should reduce false positives on real videos since misclassified frames would no longer share a single artifact-driven decision.
  • The independent-dual-decision trick could be borrowed by any binary classifier that suffers from asymmetric base rates, e.g., anomaly detection where the normal class is over-sampled, as a cheap way to avoid complementarity bias.
  • The prototype-update rule assumes the real class has meaningful subclusters; an easy test is to vary $K$ on a dataset with one dominant ethnicity or lighting condition and check whether the memory collapses.
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 / 7 minor

Summary. The paper proposes RealID, a deepfake detection method trained on FF++ only, with two new components: a Real Concept Capture Module (RealC2) that maintains multiple real-face prototypes in a memory and uses distinction and diversity losses, and an Independent Dual-Decision Classifier (IDC) that outputs four logits with a regularization term intended to decouple the real and fake probability estimates. The authors report cross-dataset AUC on Celeb-DF, DFD, DFDC, DFDCp, and UADFV, claiming an average AUC of 91.06 versus 89.32 for the strongest prior baseline SBI, a 1.74-point improvement, along with ablations and backbone-transfer experiments.

Significance. If the empirical claim were established, the work would be a useful and modular contribution to generalizable deepfake detection: the method is simple, backbone-agnostic, and the paper includes experiments on five datasets, three backbones, and module/loss ablations. The proposed mechanism of learning diverse real-face prototypes rather than only forgery artifacts is well motivated and the qualitative t-SNE and heatmap evidence supports the authors' interpretation. However, the central quantitative claim is currently not supported because the hyperparameters were selected using two of the five target test sets, and no uncertainty quantification is provided; the 1.74-point average gain could shrink or disappear under a properly held-out validation protocol and multi-seed evaluation.

major comments (3)
  1. [Section 4.1 and 4.3 (Figure 4)] The hyperparameters λ1, λ2, and λ3 in Eq. (15) are selected via grid search using performance on Celeb-DF and DFDC, which are two of the five datasets that define the headline average AUC in Table 1. The text states that "based on the combined performance of the three parameters, we empirically selected λ1 = 0.6, λ2 = 1.0, and λ3 = 1.0." This makes the reported 1.74-point average improvement over SBI an in-sample estimate on target test sets, not a cross-dataset generalization result. Please re-run the selection procedure on a validation set disjoint from all target test datasets (for example, a held-out split of FF++ or an additional dataset not used in the final evaluation), or explicitly reframe the reported numbers as test-set-tuned rather than generalization numbers.
  2. [Tables 1, 2, 3, 4] All reported AUC values are single-run numbers without error bars, standard deviations, or numbers of trials. The headline contribution is a 1.74-average-AUC gain over SBI; in deepfake detection this margin can easily be within run-to-run variance, especially with small batch sizes and random prototype initialization. The phrase "significantly outperforms" is therefore not statistically supported. Please provide at least three to five independent runs per configuration and report mean ± std, and clearly state whether baseline numbers are taken from the original papers or from re-implementations under the same protocol as RealID.
  3. [Section 3.1 and Section 4.1] Several free parameters of the method are never specified: the number of real prototypes K, the diversity margin α in Eq. (9), and the regularization intensity β in Eq. (13). Only λ1, λ2, λ3 are reported. Since these parameters control the core RealC2 and IDC mechanisms, the experiments are not fully reproducible and the sensitivity of the method to these choices is unknown. Please report all hyperparameter values and include ablations on K, α, and β (or justify why they are fixed constants).
minor comments (7)
  1. [Abstract] The abstract reports a "1.74% improvement in average accuracy," but the evaluation metric throughout the paper is AUC; please use "average AUC" for consistency and precision.
  2. [Eq. (7)] The denominator in the Prototype Distinction Loss is typeset as "2KP k=K exp(·)", which is garbled and does not parse as a sum over the K real prototypes plus the fake mean prototype; please rewrite it with proper summation notation.
  3. [Eq. (13)] The Independent Dual-Decision Regularization as written is discontinuous at |d| = 1 when β = 1.0, since the first branch gives β and the second branch gives 0 at that point; please clarify the intended functional form or correct the typo.
  4. [Figure 4 caption] The caption reads "For λ1, λ1, and λ1" where it should read "λ1, λ2, and λ3."
  5. [Section 5] The conclusion refers to "Independent Decision Classification Module" instead of the earlier "Independent Dual-Decision Classifier" (IDC); please make the terminology consistent.
  6. [Section 4.1] Training details such as optimizer, learning rate, number of epochs, number of training frames, and data augmentation are not reported; these are needed for reproducibility.
  7. [Section 4.4] The phrase "no-cherry-pick heatmaps" should be rephrased as "non-cherry-picked heatmaps" or "randomly selected heatmaps" for clarity.

Circularity Check

1 steps flagged · score 4.0 of 10

No derivation-level circularity, but λ1–λ3 are selected using Celeb-DF and DFDC, two of the five test datasets, so part of the reported cross-dataset gain is an in-sample fit.

  1. fitted input called prediction [Section 4.3, Parameter Analysis, Figure 4]
    "The performance is then evaluated on DFDC and Celeb-DF. ... Based on the combined performance of the three parameters, we empirically selected λ1 = 0.6, λ2 = 1.0, and λ3 = 1.0 as the optimal configuration."

    Celeb-DF and DFDC are two of the five datasets that define the reported average AUC in Table 1 (RealID: 95.16 on Celeb-DF, 74.67 on DFDC, average 91.06). The grid search in Section 4.3 uses the AUC on these same datasets as the selection criterion for λ1, λ2, and λ3. Consequently, the reported performance on these two datasets is an optimized in-sample value rather than an independent cross-dataset prediction. The '1.74% average improvement' headline therefore partially depends on hyperparameters chosen to maximize the evaluation metric, which is a fitted input called prediction. The other three datasets (DFD, DFDCp, UADFV) and the architecture comparisons provide independent support, so the circularity is only partial.

full rationale

The paper is an empirical systems paper; there is no formal derivation chain whose conclusion is equivalent to its premises. The self-citations (Cheng et al. 2023 and 2024a; Wang et al. 2024) appear only as related-work context and are not load-bearing. The RealC2 and IDC modules are defined by their own losses and architecture, and their contributions are supported by ablations on datasets not used for λ selection. However, the hyperparameter selection protocol compromises the out-of-sample meaning of two of the five test datasets. Because the final configuration is chosen from curves evaluated on Celeb-DF and DFDC, the reported AUCs on those datasets are in-sample and the 1.74% average gain over SBI is partly a fitted quantity. This is a partial circularity in the empirical evidence, not in the method's derivation. Score 4 reflects that the central claim retains independent content on the other datasets and in the architecture, but the headline cross-dataset average is not fully out-of-sample.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim is an empirical method claim, so the ledger is dominated by tuned hyperparameters rather than derived constants. The most important items are lambda1-lambda3, which were grid-searched with performance measured on the target test datasets, and the unreported K, alpha, and beta, which prevent exact replication. No new physical entities are introduced.

free parameters (6)
  • lambda1 = 0.6
    Weight on LDiversity; chosen by grid search with performance evaluated on the target test datasets (Figure 4).
  • lambda2 = 1.0
    Weight on LDistinction; chosen by grid search with performance evaluated on the target test datasets (Figure 4).
  • lambda3 = 1.0
    Weight on IDC regularization R; chosen by grid search with performance evaluated on the target test datasets (Figure 4).
  • Number of real prototypes K = not reported
    Controls capacity of the Multi-Real Memory; no value is given, so the method cannot be re-implemented exactly.
  • Diversity margin alpha = not reported
    Margin in LDiversity (Eq. 9); value not specified.
  • IDC regularization intensity beta = not reported
    Intensity in R (Eq. 13); value not specified.
assumptions (4)
  • domain assumption Real faces can be usefully represented by K discrete prototypes with subclasses, and forcing features toward nearest prototypes improves generalization.
    Underpins RealC2; not proven, and the optimal K is not reported.
  • ad hoc to paper Softmax over four output dimensions plus the regularization in Eq. 13 lets incorrect-class probability decrease without damaging the correct-class probability.
    The IDC mechanism is justified only by ablation results; no theoretical argument is given.
  • domain assumption Training on FF++ and evaluating on Celeb-DF, DFD, DFDC, DFDCp, and UADFV is a valid proxy for generalization.
    Standard protocol, but the hyperparameter selection uses two of these test sets, weakening the assumption.
  • standard math Standard deep learning operations (softmax, L2 normalization, cross-entropy) behave as expected.
    Used throughout Equations 1-15; no need to prove.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Real Facial Concepts for Independent Deepfake Detection." pith.science (2026). https://pith.science/paper/2F2WX52Q

@misc{pith2026250504460,
  author       = {Pith},
  title        = {Pith review of: Learning Real Facial Concepts for Independent Deepfake Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2F2WX52Q}},
  note         = {Machine review of arXiv:2505.04460}
}
read the original abstract

Deepfake detection models often struggle with generalization to unseen datasets, manifesting as misclassifying real instances as fake in target domains. This is primarily due to an overreliance on forgery artifacts and a limited understanding of real faces. To address this challenge, we propose a novel approach RealID to enhance generalization by learning a comprehensive concept of real faces while assessing the probabilities of belonging to the real and fake classes independently. RealID comprises two key modules: the Real Concept Capture Module (RealC2) and the Independent Dual-Decision Classifier (IDC). With the assistance of a MultiReal Memory, RealC2 maintains various prototypes for real faces, allowing the model to capture a comprehensive concept of real class. Meanwhile, IDC redefines the classification strategy by making independent decisions based on the concept of the real class and the presence of forgery artifacts. Through the combined effect of the above modules, the influence of forgery-irrelevant patterns is alleviated, and extensive experiments on five widely used datasets demonstrate that RealID significantly outperforms existing state-of-the-art methods, achieving a 1.74% improvement in average accuracy.

Figures

Figures reproduced from arXiv: 2505.04460 by the authors.

Figure 1
Figure 1. (a) The real samples are misclassified as fake (highlighted [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overall architecture of the RealID framework. RealID consists of two main modules: (i) the Real Concept Capture( [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the update process for real facial prototypes. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: AUC (%) comparison under different hyperparameter [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Illustration for no-cherry-pick heatmaps from four [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Pinco: Position-induced Consistent Adapter for Diffusion Transformer in Foreground-conditioned Inpainting

    cs.CV 2024-12 conditional novelty 6.0 of 10

    Pinco is a plug-and-play adapter that enables diffusion transformers to inpaint backgrounds around a provided foreground object, preserving its shape via self-attention injection and a positional anchor.

  2. Fair Deepfake Detectors Can Generalize

    cs.LG 2025-07 reject novelty 5.0 of 10

    The paper argues that demographic fairness interventions can causally improve cross-domain generalization in deepfake detection and introduces DAID to achieve both, but the causal evidence is flawed.

Reference graph

Works this paper leans on

51 extracted references · 50 canonical work pages · cited by 2 Pith papers

  1. [1]

    Towards open-set identity preserving face synthesis

    Jianmin Bao, Dong Chen, Fang Wen, Houqiang Li, and Gang Hua. Towards open-set identity preserving face synthesis. In CVPR , pages 6713--6722, 2018

    work page 2018

  2. [2]

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

    Junyi Cao, Chao Ma, Taiping Yao, Shen Chen, Shouhong Ding, and Xiaokang Yang. End-to-end reconstruction-classification learning for face forgery detection. In CVPR , pages 4103--4112, 2022

    work page 2022

  3. [3]

    Self-supervised learning of adversarial example: Towards good generalizations for deepfake detection

    Liang Chen, Yong Zhang, Yibing Song, Lingqiao Liu, and Jue Wang. Self-supervised learning of adversarial example: Towards good generalizations for deepfake detection. In CVPR , pages 18689--18698, 2022

    work page 2022

  4. [4]

    Voice-face homogeneity tells deepfake

    Harry Cheng, Yangyang Guo, Tianyi Wang, Qi Li, Xiaojun Chang, and Liqiang Nie. Voice-face homogeneity tells deepfake. ACM TOMM , pages 1--22, 2023

    work page 2023

  5. [5]

    Diffusion facial forgery detection

    Harry Cheng, Yangyang Guo, Tianyi Wang, Liqiang Nie, and Mohan Kankanhalli. Diffusion facial forgery detection. In ACM MM , page 5939–5948, 2024

    work page 2024

  6. [6]

    Can we leave deepfake data behind in training deepfake detector? In NeurIPS , 2024

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

    work page 2024

  7. [7]

    Exploiting style latent flows for generalizing deepfake video detection

    Jongwook Choi, Taehoon Kim, Yonghyun Jeong, Seungryul Baek, and Jongwon Choi. Exploiting style latent flows for generalizing deepfake video detection. In CVPR , pages 1133--1143, 2024

    work page 2024

  8. [8]

    The deepfake detection challenge dataset

    Brian Dolhansky, Joanna Bitton, Ben Pflaum, Jikuo Lu, Russ Howes, Menglin Wang, and Cristian Canton - Ferrer. The deepfake detection challenge dataset. CoRR , pages 1--13, 2020

    work page 2020

Show all 51 references
  1. [9]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at...

    2021

  2. [10]

    Deepfake detection dataset, 2020

    Nick Dufour and Andrew Gully. Deepfake detection dataset, 2020

    2020

  3. [11]

    Learning second order local anomaly for general face forgery detection

    Jianwei Fei, Yunshu Dai, Peipeng Yu, Tianrun Shen, Zhihua Xia, and Jian Weng. Learning second order local anomaly for general face forgery detection. In CVPR , pages 20238--20248, 2022

    2022

  4. [12]

    Goodfellow, Jean Pouget - Abadie, Mehdi Mirza, Bing Xu, David Warde - Farley, Sherjil Ozair, Aaron C

    Ian J. Goodfellow, Jean Pouget - Abadie, Mehdi Mirza, Bing Xu, David Warde - Farley, Sherjil Ozair, Aaron C. Courville, and Yoshua Bengio. Generative adversarial nets. In NIPS , pages 2672--2680, 2014

    2014

  5. [13]

    Improving generalization of deepfake detectors by imposing gradient regularization

    Weinan Guan, Wei Wang, Jing Dong, and Bo Peng. Improving generalization of deepfake detectors by imposing gradient regularization. IEEE TIFS , 19:5345--5356, 2024

    2024

  6. [14]

    Leveraging real talking faces via self-supervision for robust forgery detection

    Alexandros Haliassos, Rodrigo Mira, Stavros Petridis, and Maja Pantic. Leveraging real talking faces via self-supervision for robust forgery detection. In CVPR , pages 14930--14942, 2022

    2022

  7. [15]

    Contrastive learning for deepfake classification and localization via multi-label ranking

    Cheng - Yao Hong, Yen - Chi Hsu, and Tyng - Luh Liu. Contrastive learning for deepfake classification and localization via multi-label ranking. In CVPR , pages 17627--17637, 2024

    2024

  8. [16]

    Exploring frequency adversarial attacks for face forgery detection

    Shuai Jia, Chao Ma, Taiping Yao, Bangjie Yin, Shouhong Ding, and Xiaokang Yang. Exploring frequency adversarial attacks for face forgery detection. In CVPR , pages 4093--4102, 2022

    2022

  9. [17]

    A style-based generator architecture for generative adversarial networks

    Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. In CVPR , pages 4401--4410, 2019

    2019

  10. [18]

    Hasam Khalid, Shahroz Tariq, Minha Kim, and Simon S. Woo. Fakeavceleb: A novel audio-video multimodal deepfake dataset. In NIPS , pages 1--15, 2021

    2021

  11. [19]

    Kingma and Max Welling

    Diederik P. Kingma and Max Welling. Auto-encoding variational bayes. In ICLR , pages 1--14, 2014

    2014

  12. [20]

    Kodf: A large-scale korean deepfake detection dataset

    Patrick Kwon, Jaeseong You, Gyuhyeon Nam, Sungwoo Park, and Gyeongsu Chae. Kodf: A large-scale korean deepfake detection dataset. In ICCV , pages 10724--10733, 2021

    2021

  13. [21]

    Nguyen, Junichi Yamagishi, and Isao Echizen

    Trung - Nghia Le, Huy H. Nguyen, Junichi Yamagishi, and Isao Echizen. Openforensics: Large-scale challenging dataset for multi-face forgery detection and segmentation in-the-wild. In ICCV , pages 10097--10107, 2021

    2021

  14. [22]

    In ictu oculi: Exposing ai generated fake face videos by detecting eye blinking

    Yuezun Li, Ming-Ching Chang, and Siwei Lyu. In ictu oculi: Exposing ai generated fake face videos by detecting eye blinking. In WIFS , 2018

    2018

  15. [23]

    Face x-ray for more general face forgery detection

    Lingzhi Li, Jianmin Bao, Ting Zhang, Hao Yang, Dong Chen, Fang Wen, and Baining Guo. Face x-ray for more general face forgery detection. In CVPR , pages 5000--5009, 2020

    2020

  16. [24]

    Celeb-df: A large-scale challenging dataset for deepfake forensics

    Yuezun Li, Xin Yang, Pu Sun, Honggang Qi, and Siwei Lyu. Celeb-df: A large-scale challenging dataset for deepfake forensics. In CVPR , pages 3204--3213, 2020

    2020

  17. [25]

    Logical relation inference and multiview information interaction for domain adaptation person re-identification

    Shuang Li, Fan Li, Jinxing Li, Huafeng Li, Bob Zhang, Dapeng Tao, and Xinbo Gao. Logical relation inference and multiview information interaction for domain adaptation person re-identification. IEEE TNNLS , 2023

    2023

  18. [26]

    Generalizing face forgery detection with high-frequency features

    Yuchen Luo, Yong Zhang, Junchi Yan, and Wei Liu. Generalizing face forgery detection with high-frequency features. In CVPR , pages 16317--16326, 2021

    2021

  19. [27]

    Two-branch recurrent network for isolating deepfakes in videos

    Iacopo Masi, Aditya Killekar, Royston Marian Mascarenhas, Shenoy Pratik Gurudatt, and Wael AbdAlmageed. Two-branch recurrent network for isolating deepfakes in videos. In ECCV , pages 667--684, 2020

    2020

  20. [28]

    Df-platter: Multi-face heterogeneous deepfake dataset

    Kartik Narayan, Harsh Agarwal, Kartik Thakral, Surbhi Mittal, Mayank Vatsa, and Richa Singh. Df-platter: Multi-face heterogeneous deepfake dataset. In CVPR , pages 9739--9748, 2023

    2023

  21. [29]

    CORE: consistent representation learning for face forgery detection

    Yunsheng Ni, Depu Meng, Changqian Yu, Chengbin Quan, Dongchun Ren, and Youjian Zhao. CORE: consistent representation learning for face forgery detection. In CVPRW , pages 12--21, 2022

    2022

  22. [30]

    Deepfake detection based on discrepancies between faces and their context

    Yuval Nirkin, Lior Wolf, Yosi Keller, and Tal Hassner. Deepfake detection based on discrepancies between faces and their context. IEEE TPAMI , 44(10):6111--6121, 2022

    2022

  23. [31]

    Learning memory-guided normality for anomaly detection

    Hyunjong Park, Jongyoun Noh, and Bumsub Ham. Learning memory-guided normality for anomaly detection. In CVPR , pages 14372--14381, 2020

    2020

  24. [32]

    Fully unsupervised deepfake video detection via enhanced contrastive learning

    Tong Qiao, Shichuang Xie, Yanli Chen, Florent Retraint, and Xiangyang Luo. Fully unsupervised deepfake video detection via enhanced contrastive learning. IEEE TPAMI , 46(7):4654--4668, 2024

    2024

  25. [33]

    Faceforensics++: Learning to detect manipulated facial images

    Andreas R \" o ssler, Davide Cozzolino, Luisa Verdoliva, Christian Riess, Justus Thies, and Matthias Nie ner. Faceforensics++: Learning to detect manipulated facial images. In ICCV , pages 1--11, 2019

    2019

  26. [34]

    Detecting deepfakes with self-blended images

    Kaede Shiohara and Toshihiko Yamasaki. Detecting deepfakes with self-blended images. In CVPR , pages 18699--18708, 2022

    2022

  27. [35]

    Rethinking open-world deepfake attribution with multi-perspective sensory learning

    Zhimin Sun, Shen Chen, Taiping Yao, Ran Yi, Shouhong Ding, and Lizhuang Ma. Rethinking open-world deepfake attribution with multi-perspective sensory learning. IJCV , pages 1--24, 2024

    2024

  28. [36]

    Seitz, and Ira Kemelmacher - Shlizerman

    Supasorn Suwajanakorn, Steven M. Seitz, and Ira Kemelmacher - Shlizerman. Synthesizing obama: learning lip sync from audio. ACM TOG , 36(4):95:1--95:13, 2017

    2017

  29. [37]

    Mingxing Tan and Quoc V. Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In ICML , volume 97, pages 6105--6114, 2019

    2019

  30. [38]

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

    Chuangchuang Tan, Huan Liu, 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 CVPR , pages 28130--28139, 2024

    2024

  31. [39]

    Face2face: Real-time face capture and reenactment of RGB videos

    Justus Thies, Michael Zollh \" o fer, Marc Stamminger, Christian Theobalt, and Matthias Nie ner. Face2face: Real-time face capture and reenactment of RGB videos. In CVPR , pages 2387--2395, 2016

    2016

  32. [40]

    Learning to discover forgery cues for face forgery detection

    Jiahe Tian, Peng Chen, Cai Yu, Xiaomeng Fu, Xi Wang, Jiao Dai, and Jizhong Han. Learning to discover forgery cues for face forgery detection. IEEE TIFS , 19:3814--3828, 2024

    2024

  33. [41]

    Representative forgery mining for fake face detection

    Chengrui Wang and Weihong Deng. Representative forgery mining for fake face detection. In CVPR , pages 14923--14932, 2021

    2021

  34. [42]

    Deepfake detection: A comprehensive survey from the reliability perspective

    Tianyi Wang, Xin Liao, Kam Pui Chow, Xiaodong Lin, and Yinglong Wang. Deepfake detection: A comprehensive survey from the reliability perspective. ACM CSUR , 57(3), 2024

    2024

  35. [43]

    Sstnet: Detecting manipulated faces through spatial, steganalysis and temporal features

    Xi Wu, Zhen Xie, YuTao Gao, and Yu Xiao. Sstnet: Detecting manipulated faces through spatial, steganalysis and temporal features. In ICASSP , pages 2952--2956, 2020

    2020

  36. [44]

    Mmnet: multi-collaboration and multi-supervision network for sequential deepfake detection

    Ruiyang Xia, Decheng Liu, Jie Li, Lin Yuan, Nannan Wang, and Xinbo Gao. Mmnet: multi-collaboration and multi-supervision network for sequential deepfake detection. IEEE TIFS , 2024

    2024

  37. [45]

    Advancing generalized deepfake detector with forgery perception guidance

    Ruiyang Xia, Dawei Zhou, Decheng Liu, Lin Yuan, Shuodi Wang, Jie Li, Nannan Wang, and Xinbo Gao. Advancing generalized deepfake detector with forgery perception guidance. In ACM MM , pages 6676--6685, 2024

    2024

  38. [46]

    Region-aware face swapping

    Chao Xu, Jiangning Zhang, Miao Hua, Qian He, Zili Yi, and Yong Liu. Region-aware face swapping. In CVPR , pages 7622--7631, 2022

    2022

  39. [47]

    Text-guided human image manipulation via image-text shared space

    Xiaogang Xu, Ying - Cong Chen, Xin Tao, and Jiaya Jia. Text-guided human image manipulation via image-text shared space. IEEE TPAMI , 44(10):6486--6500, 2022

    2022

  40. [48]

    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. In ICCV , pages 22355--22366, 2023

    2023

  41. [49]

    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

  42. [50]

    Multi-attentional deepfake detection

    Hanqing Zhao, Wenbo Zhou, Dongdong Chen, Tianyi Wei, Weiming Zhang, and Nenghai Yu. Multi-attentional deepfake detection. In CVPR , pages 2185--2194, 2021

    2021

  43. [51]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence '...

Pith tools

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