The DOI matches its Crossref record; 60 of the 60 references it cites were found in a registry or an index; 19 of the 30 sources its claims cite were checked for support.
Cross-dataset video deepfake detection using Transformer and CNN architectures
Open accessAbstract
The growing sophistication of deepfake generation techniques poses serious challenges to the authenticity of digital media, with potential risks spanning privacy, security and misinformation. Deep learning (DL) methods have shown significant promise in detecting such manipulations; however, inconsistencies in application, the absence of standardized pipelines and limited cross-dataset generalization hinder their reliable deployment in real-world scenarios. This work presents a comprehensive evaluation of Transformer- and CNN-based architectures for video deepfake detection. Multiple benchmark datasets, along with a novel facial-reenactment dataset, are used to investigate cross-dataset generalization and pretraining with limited fine-tuning on small target subsets (10–30%). Additionally, we analyze the impact of temporal window length on detection performance. Experimental results demonstrate that TimeSformer consistently achieves the highest performance, reaching 78.4% accuracy, 0.801 area under the curve (AUC) and 77.0% F1-score with 96-frame clips and 30% fine-tuning, confirming the advantage of joint spatiotemporal modeling. All models benefit from moderate fine-tuning, with gains plateauing beyond 20%. Increasing clip length enhances performance for temporally aware models, highlighting the importance of extended temporal context. Overall, this study provides empirical evidence into the strengths and limitations of current architectures, offering guidance for future research and practical deployment of robust and generalizable deepfake detectors.
Overview
CoChat Summary
TL;DR
This paper compares CNN- and Transformer-based video deepfake detectors in a cross-dataset setting, using FF++ and Celeb-DF for training and DFD plus ReenactFaces for testing. TimeSformer performs best overall, reaching 78.4% accuracy, 0.801 AUC, and 77.0% F1-score with 96-frame clips and 30% fine-tuning.
Key findings
- Across all configurations, TimeSformer achieved the strongest results; in the 96-frame, 30% fine-tuning setting it reached 78.4% accuracy / 0.801 AUC / 77.0% F1-score.
- Fine-tuning improved every model, with the paper stating the largest gains were between 0% and 20% adaptation and that performance then tended to plateau.
- Longer temporal windows generally helped temporally aware models: the paper reports better performance for 32, 64, and 96-frame clips in CNN-LSTM, CNN-Attention, and TimeSformer, while ResNet models improved only modestly.
- CNN-Attention outperformed the ResNet variants and CNN-LSTM; for example at 96 frames and 30% fine-tuning, it reached 75.6% accuracy / 0.752 AUC / 73.6% F1-score, compared with 72.8% / 0.744 / 70.9% for ResNet101.
Methods
- Cross-dataset evaluation with two source datasets (FF++ and Celeb-DF) and two target datasets (DFD and ReenactFaces), testing both direct transfer (0% fine-tuning) and partial adaptation (10%, 20%, 30% fine-tuning).
- Datasets were preprocessed by frame extraction, RetinaFace-based face detection/alignment, resizing to 112 × 112, normalization, and organization into temporal windows of 32, 64, or 96 frames.
- Compared six architectures: ResNet50, ResNet101, CNN-LSTM, CNN-Attention, ViT, and TimeSformer; models were trained with binary cross-entropy and Adam (1 × 10^-4), evaluated with accuracy, AUC, and F1-score, and results were averaged across three runs.
Limitations
- The paper states that the datasets may not fully represent emerging generative techniques, particularly diffusion or neural rendering models.
- The experimental setup uses visual information alone, without audio or physiological signals, which the paper says could strengthen detection in harder cases.
- The paper notes that it did not include explicit domain adaptation modules and did not include some task-specific Transformer methods (e.g., GenConViT, DeepFake-Adapter, Swin-Fake), which may limit robustness; this is a limitation stated in the paper.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
13 Valid · 3 Not valid · 5 Not checked · 47 Own results and interpretations
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“the absence of standardized pipe -lines makes it difficult to compare methods or deploy them reliably in real-world scenarios”
What the sources say
- Could not verifycites [5] Mustak, M., Salminen, J., Mäntymäki, M., Rahman, A., Dwivedi, Y.K.: Deepfakes: deceptions, mitigations, and opportunities. J. Bus. Res. 154, 113368 (2023). https://doi.org/10.1016/j.jbusres.2022.113368 sourceExperimental — accuracy not yet measured
Section 4.2.1 discusses limitations of deepfake detection and the need for successful deployment, but attributes those limitations to rapid advances in synthetic media. Our copy does not address whether a lack of standardized pipelines makes methods difficult to compare or deploy reliably (searched: standardized pipelines, benchmarks, and real-world deployment). (LLM, conf=0.74)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Early deepfake detection methods primarily relied on conventional CNNs to identify low-level visual inconsistencies and blending artifacts in manipulated frames.”
What the sources say
- Partly supportedcites [9] Afchar, D., Nozick, V., Yamagishi, J., Echizen, I.: MesoNet: a compact facial video forgery detection network. In: 2018 IEEE International Workshop on Information Forensics and Security (WIFS), 1–7. (2018) sourceExperimental — accuracy not yet measured
The Proposed method (p. 3) describes compact convolutional networks for detecting forged facial videos, and §3.6 (p. 6) links their responses to differences in facial detail and blurriness. The paper does not establish that early methods primarily relied on CNNs, and it describes its own analysis as mesoscopic rather than low-level; it does not show that the networks specifically detect blending artifacts. (LLM, conf=0.86)
- Partly supportedcites [10] Rössler, A., Cozzolino, D., Verdoliva, L., Rieß, C., Thies, J., Nießner, M.: FaceForensics++: learning to detect manipulated facial images. 2019 IEEE/CVF Int. Conf. Comput. Vis. (ICCV). (2019). https://doi.org/10.1109/iccv.2019.00009 sourceExperimental — accuracy not yet measured
The Face Manipulation Methods section says some detectors use deep networks to capture subtle low- or high-level inconsistencies, supporting that part of the claim. It does not say early deepfake detectors primarily used conventional CNNs or specifically targeted blending artifacts; its broader multimedia-forensics discussion instead describes early methods as handcrafted-feature approaches and CNN-based work as more recent. Page numbers are unavailable in our copy. (LLM, conf=0.84)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Studies such as Li et al. [25] and Agarwal et al. [26] demonstrated that most detectors trained on one dataset (i.e., FF++) exhibit significant performance drops when tested on others (i.e., Celeb-DF, DFDC).”
What the sources say
- Partly supportedcites [25] Li, Y., Yang, X., Sun, P., Qi, H., Lyu, S.: Celeb-DF: A large-scale challenging dataset for deepfake forensics. 2020 IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). (2020). https://doi.org/10.1109/cvpr42600.2020.00327 sourceExperimental — accuracy not yet measured
The Experimental Settings (p. 7) identify Xception as trained on FaceForensics++, and the Results table (p. 8) shows its AUC dropping sharply from FF-DF to DFDC and Celeb-DF. The Results and Analysis (p. 7) also report lower performance across methods on those newer datasets. But the authors used published pretrained models with varied training datasets; they do not establish that most of the evaluated detectors were trained on FF++. (LLM, conf=0.88)
- Not matchedcites [26] Agarwal, S., Farid, H., Gu, Y., He, M., Nagano, K., Li, H.: Protecting world leaders against deep fakes. In: CVPR Workshops (2019)
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
boostedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Vision Transformer (ViT) [17], and video-specific variants such as TimeSformer [ 18, 19 ] and Video Swin Transformer [20] capture longrange dependencies and global spatiotemporal context that CNNs often miss.”
What the sources say
- Partly supportedcites [17] Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., Houlsby, N.: An image is worth 16x16 words: transformers for image recognition at scale. In: International Conference on Learning Representations. (2021). https://openreview.net/pdf?id=YicbFdNTTy sourceExperimental — accuracy not yet measured
Section 4.5 (p. 8) reports that ViT integrates information across the entire image, including in its lowest layers; Section 3.1 (p. 4) contrasts ViT’s global self-attention with locality built into CNNs. This paper studies image recognition, not video, so it does not support the claim about temporal context or the named video variants. It also does not establish that CNNs often miss such context. (LLM, conf=0.87)
- Partly supportedcites [18] Bertasius, G., Wang, H., Torresani, L.: Is Space-time attention all you need for video understanding? In: International Conference on Machine Learning, 813–824. (2021). http://proceedings.mlr.press/v139/bertasius21a/bertasius21a.pdf sourceExperimental — accuracy not yet measured
The TimeSformer results (Test, page not provided) report that the model exploits long-range temporal dependencies, and its space-time attention visualizations show attention to relevant video regions. The cited paper does not establish that CNNs often miss global spatiotemporal context; it also does not evaluate Video Swin Transformer. (LLM, conf=0.84)
- Supportedcites [19] Petmezas, G., Papageorgiou, V.E., Vassilikos, V., Pagourelias, E., Tachmatzidis, D., Tsaklidis, G., Katsaggelos, A.K., Maglaveras, N.: Automated HFREF diagnosis using an optimized TimeSformer model in echocardiography. J. Imaging Inf. Med. (2025). https://doi.org/10.1007/s10278-025-01758-2 source
The citation to this paper is attached to TimeSformer, not to ViT or Video Swin Transformer. Materials and Methods says TimeSformer uses spatial and temporal self-attention to capture complex video dependencies, contrasting it with limitations of CNN-based methods; the Discussion likewise reports that it learned complex spatiotemporal features. (LLM, conf=0.93)
- Not checked — no full text of the sourcecites [20] Liu, Z., Ning, J., Cao, Y., Wei, Y., Zhang, Z., Lin, S., Hu, H.: Video swin transformer. 2022 IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 3192, 3201 (2022). https://doi.org/10.1109/cvpr52688.2022.00320 source
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“FF++ [ 10] is one of the most widely used deepfake datasets, containing 1000 pristine videos collected from YouTube and manipulated using four prominent synthesis techniques”
What the sources say
- Supportedcites [10] Rössler, A., Cozzolino, D., Verdoliva, L., Rieß, C., Thies, J., Nießner, M.: FaceForensics++: learning to detect manipulated facial images. 2019 IEEE/CVF Int. Conf. Comput. Vis. (ICCV). (2019). https://doi.org/10.1109/iccv.2019.00009 source
Section 3, “Large-Scale Facial Forgery Database” (page not provided), states that four face-manipulation methods were applied to 1,000 pristine videos and that the videos were collected from YouTube. This supports the attributed description of FF++. (LLM, conf=0.95)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
21 more claims
Create a free account to see all 68 claims and every source verdict.
- It may state two separate claims as one. 30
- Probably not a substantial claim. 14
- The quote only partly states this claim. 1
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 4 Results
“The TimeSformer achieved its best result at 30% fine-tuning with 96-frame clips (78.4% accuracy, 0.801 AUC, 70.0% F1-score), representing the optimal trade-off between adaptation and generalization.”
§ 4 Results
“TimeSformer (F) | 71.2 / 0.722 / 69.6 | 77.3 / 0.769 / 75.4 | 77.5 / 0.789 / 75.9 | 78.4 / 0.801 / 77.0 |”
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472
Wang, X., Girshick, R., Gupta, A., & He, K. (2017) · arXiv · Cited by 11,689
Wang, X., Girshick, R., Gupta, A., & He, K. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00813 · Cited by 11,596 · cited 1×
“Unlike CNN–LSTM, this design avoids recurrent processing and enables parallel computation across frames, improving training efficiency while retaining sensitivity to temporally salient forgery patterns [41].”
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2015) · arXiv · doi:10.48550/arxiv.1512.03385 · Cited by 4,196
FaceForensics++: Learning to Detect Manipulated Facial Images
Rossler, A., Cozzolino, D., Verdoliva, L., Riess, C., Thies, J., & Niessner, M. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00009 · Cited by 3,314
Liu, Z., Ning, J., Cao, Y., Wei, Y., Zhang, Z., Lin, S., & Hu, H. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00320 · Cited by 2,294
MesoNet: a Compact Facial Video Forgery Detection Network
Afchar, D., Nozick, V., Yamagishi, J., & Echizen, I. (2018) · International Workshop on Information Forensics and Security · doi:10.48550/arxiv.1809.00888 · Cited by 1,940
MesoNet: a Compact Facial Video Forgery Detection Network
Afchar, D., Nozick, V., Yamagishi, J., & Echizen, I. (2018) · 2018 IEEE International Workshop on Information Forensics and Security (WIFS) · doi:10.1109/wifs.2018.8630761 · Cited by 1,872
Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Forensics
Li, Y., Yang, X., Sun, P., Qi, H., & Lyu, S. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00327 · Cited by 1,846 · cited 2×
“Celeb-DF [ 25] extends the evaluation to more challenging and realistic manipulations.”
RetinaFace: Single-Shot Multi-Level Face Localisation in the Wild
Deng, J., Guo, J., Ververas, E., Kotsia, I., & Zafeiriou, S. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00525 · Cited by 1,681
Is Space-Time Attention All You Need for Video Understanding?
Bertasius, G., Wang, H., & Torresani, L. (2021) · arXiv · doi:10.48550/arxiv.2102.05095 · Cited by 1,663
Face X-Ray for More General Face Forgery Detection
Li, L., Bao, J., Zhang, T., Yang, H., Chen, D., Wen, F., & Guo, B. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00505 · Cited by 1,211
Thinking in Frequency: Face Forgery Detection by Mining Frequency-Aware Clues
Qian, Y., Yin, G., Sheng, L., Chen, Z., & Shao, J. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58610-2_6 · Cited by 1,058
Media Forensics and DeepFakes: An Overview
Verdoliva, L. (2020) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2020.3002101 · Cited by 979
Capsule-forensics: Using Capsule Networks to Detect Forged Images and Videos
Nguyen, H. H., Yamagishi, J., & Echizen, I. (2019) · ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2019.8682602 · Cited by 951
Deepfake Detection: A Systematic Literature Review
Rana, M. S., Nobi, M. N., Murali, B., & Sung, A. H. (2022) · IEEE Access · doi:10.1109/access.2022.3154404 · Cited by 573
Lips Don't Lie: A Generalisable and Robust Approach to Face Forgery Detection
Haliassos, A., Vougioukas, K., Petridis, S., & Pantic, M. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00500 · Cited by 537
Deepfake detection using deep learning methods: A systematic and comprehensive review
Heidari, A., Jafari Navimipour, N., Dag, H., & Unal, M. (2024) · WIREs Data Mining and Knowledge Discovery · doi:10.1002/widm.1520 · Cited by 349 · cited 1×
“A summary of the datasets used in this study, along with their key characteristics and manipulation types, is provided in Table 1. discriminative patterns related to texture, illumination and blending inconsistencies often present in manipulated facial regions [37].”
Automated Atrial Fibrillation Detection using a Hybrid CNN-LSTM Network on Imbalanced ECG Datasets
Petmezas, G., Haris, K., Stefanopoulos, L., Kilintzis, V., Tzavelis, A., Rogers, J. A., Katsaggelos, A. K., & Maglaveras, N. (2021) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2020.102194 · Cited by 333
A Survey on Deepfake Video Detection
Yu, P., Xia, Z., Fei, J., & Lu, Y. (2021) · IET Biometrics · doi:10.1049/bme2.12031 · Cited by 253
Motion Representations for Articulated Animation
Siarohin, A., Woodford, O. J., Ren, J., Chai, M., & Tulyakov, S. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.01344 · Cited by 245
Thin-Plate Spline Motion Model for Image Animation
Zhao, J., & Zhang, H. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00364 · Cited by 222
ForensicTransfer: Weakly-supervised Domain Adaptation for Forgery Detection
Cozzolino, D., Thies, J., Rössler, A., Riess, C., Nießner, M., & Verdoliva, L. (2018) · arXiv · doi:10.48550/arxiv.1812.02510 · Cited by 221
Deepfake video detection: challenges and opportunities
Kaur, A., Noori Hoshyar, A., Saikrishna, V., Firmin, S., & Xia, F. (2024) · Artificial Intelligence Review · doi:10.1007/s10462-024-10810-6 · Cited by 214
Deepfakes: Deceptions, mitigations, and opportunities
Mustak, M., Salminen, J., Mäntymäki, M., Rahman, A., & Dwivedi, Y. K. (2023) · Journal of Business Research · doi:10.1016/j.jbusres.2022.113368 · Cited by 209
Automated Lung Sound Classification Using a Hybrid CNN-LSTM Network and Focal Loss Function
Petmezas, G., Cheimariotis, G.-A., Stefanopoulos, L., Rocha, B., Paiva, R. P., Katsaggelos, A. K., & Maglaveras, N. (2022) · Sensors · doi:10.3390/s22031232 · Cited by 200
Deepfake Attacks: Generation, Detection, Datasets, Challenges, and Research Directions
Naitali, A., Ridouani, M., Salahdine, F., & Kaabouch, N. (2023) · Computers · doi:10.3390/computers12100216 · Cited by 117
Real-Time Deepfake Video Detection Using Eye Movement Analysis with a Hybrid Deep Learning Approach
Javed, M., Zhang, Z., Dahri, F. H., & Laghari, A. A. (2024) · Electronics · doi:10.3390/electronics13152947 · Cited by 111
Video deepfake detection using a hybrid CNN-LSTM-Transformer model for identity verification
Πετμεζάς, Γ., Vanian, V., Konstantoudakis, K., Almaloglou, E. E. I., & Zarpalas, D. (2025) · Multimedia Tools and Applications · doi:10.1007/s11042-024-20548-6 · Cited by 90
Transformers for Vision: A Survey on Innovative Methods for Computer Vision
Palanisamy, B., Hassija, V., Chatterjee, A., Mandal, A., Chakraborty, D., Pandey, A., Chalapathi, G. S. S., & Kumar, D. (2025) · IEEE Access · doi:10.1109/access.2025.3571735 · Cited by 83
Enhancing multimodal deepfake detection with local–global feature integration and diffusion models
Javed, M., Zhang, Z., Dahri, F. H., & Kumar, T. (2025) · Signal Image and Video Processing · doi:10.1007/s11760-025-03970-7 · Cited by 78
Detecting Deepfake Videos with Temporal Dropout 3DCNN
Zhang, D., Li, C., Lin, F., Zeng, D., & Ge, S. (2021) · Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2021/178 · Cited by 77 · cited 1×
“Later approaches extended CNNs to incorporate temporal information, such as CNN-LSTM and 3D CNN frameworks, to capture motion artifacts and frame continuity disrupted by deepfake generation [11, 12].”
Audio–Visual Synchronization and Lip Movement Analysis for Real-Time Deepfake Detection
Javed, M., Zhang, Z., Dahri, F. H., Laghari, A. A., Krajčík, M., & Almadhor, A. S. (2025) · International Journal of Computational Intelligence Systems · doi:10.1007/s44196-025-00911-7 · Cited by 62
Deepfake video detection methods, approaches, and challenges
Alrashoud, M. (2025) · Alexandria Engineering Journal · doi:10.1016/j.aej.2025.04.007 · Cited by 49
DeepFake-Adapter: Dual-Level Adapter for DeepFake Detection
Shao, R., Wu, T., Nie, L., & Liu, Z. (2025) · International Journal of Computer Vision · doi:10.1007/s11263-024-02274-6 · Cited by 48 · cited 1×
“Similarly, Shao et al. [22 ] proposed DeepFake-Adapter, a parameter-efficient framework that augments pre-trained ViTs with dual-level adapters to enhance transferability and domain generalization.”
GenConViT: Deepfake Video Detection Using Generative Convolutional Vision Transformer
Deressa, D. W., Mareen, H., Lambert, P., Atnafu, S., Akhtar, Z., & Van Wallendael, G. (2025) · Applied Sciences · doi:10.3390/app15126622 · Cited by 46
Visual Deepfake Detection: Review of Techniques, Tools, Limitations, and Future Prospects
Ur Rehman Ahmed, N., Badshah, A., Adeel, H., Tajammul, A., Daud, A., & Alsahfi, T. (2025) · IEEE Access · doi:10.1109/access.2024.3523288 · Cited by 45
Rochow, A., Schwarz, M., & Behnke, S. (2024) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52733.2024.00737 · Cited by 43
Real-Time Deepfake Detection via Gaze and Blink Patterns: A Transformer Framework
Javed, M., Zhang, Z., Dahri, F. H., Laghari, A. A., Krajčík, M., & Almadhor, A. S. (2025) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2025.062954 · Cited by 36
Facing Completely Occluded Short-Term Tracking Based on Correlation Filters
Zhang, Y., Pan, H., Wang, J., & Sun, W. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3317483 · Cited by 32
Self-supervised Transformer for Deepfake Detection
Zhao, H., Zhou, W., Chen, D., Zhang, W., & Yu, N. (2022) · arXiv · doi:10.48550/arxiv.2203.01265 · Cited by 24
Optimizing intrinsic representation for tracking
Zhang, Y., & Sun, H. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111955 · Cited by 23
Wu, Y., Rocha, B. M., Kaimakamis, E., Cheimariotis, G.-A., Petmezas, G., Chatzis, E., Kilintzis, V., Stefanopoulos, L., Pessoa, D., Marques, A., et al. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.121089 · Cited by 23
Enabling deformation slack in tracking with temporally even correlation filters
Zhang, Y., Pan, H., & Wang, J. (2024) · Neural Networks · doi:10.1016/j.neunet.2024.106839 · Cited by 22
Swin-Fake: A Consistency Learning Transformer-Based Deepfake Video Detector
Gong, L. Y., Li, X. J., & Chong, P. H. J. (2024) · Electronics · doi:10.3390/electronics13153045 · Cited by 22 · cited 1×
“Other recent approaches, such as MSIDSnet [23] and Swin-Fake [ 24], employ dual-stream and consistency-learning strategies to fuse spatial, frequency and temporal features within Transformer architectures, yielding notable gains in generalization across manipulation types.”
Πετμεζάς, Γ., Papageorgiou, V. Ε., Vassilikos, V. P., Pagourelias, E. D., Tachmatzidis, D., Tsaklidis, G. M., Katsaggelos, A. K., & Maglaveras, N. (2025) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2025.104800 · Cited by 20
Improved Deepfake Video Detection Using Convolutional Vision Transformer
Deressa, D. W., Lambert, P., Van Wallendael, G., Atnafu, S., & Mareen, H. (2024) · 2024 IEEE Gaming, Entertainment, and Media Conference (GEM) · doi:10.1109/gem61861.2024.10585593 · Cited by 20
DeepFake detection method based on multi-scale interactive dual-stream network
Cheng, Z., Wang, Y., Wan, Y., & Jiang, C. (2024) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2024.104263 · Cited by 19
Learning Distance Constrained Transformation for Video Tracking in Car-Following
Sun, H., Zhang, Y., Zhang, H., Qiu, X., & Rudas, I. J. (2025) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2025.3574326 · Cited by 18
Decoding split-frequency representation for cross-scale tracking
Zhang, Y., & Sun, H. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.107587 · Cited by 16
Uncertainty CNNs: A path to enhanced medical image classification performance
Papageorgiou, V. Ε., Πετμεζάς, Γ., Dogoulis, P., Cordy, M., & Maglaveras, N. (2025) · Mathematical Biosciences & Engineering · doi:10.3934/mbe.2025020 · Cited by 13
A Dual-Branch Fusion Model for Deepfake Detection Using Video Frames and Microexpression Features
Petmezas, G., Vanian, V., Rufete, M. P., Almaloglou, E. E. I., & Zarpalas, D. (2025) · Journal of Imaging · doi:10.3390/jimaging11070231 · Cited by 8 · cited 1×
“Combining traditional RGB video streams with subtle microexpression or physiological features could offer a more comprehensive understanding of facial behavior, helping models distinguish between authentic and manipulated content in challenging conditions [46].”
Chen, N., Xu, T., Sun, M., Yao, C., & Yang, D. (2025) · Intelligent Computing · doi:10.34133/icomputing.0143 · Cited by 5 · cited 1×
“…exploited in contexts ranging from political disinformation campaigns to identity theft and personal harassment, highlighting the urgent need for effective detection mechanisms. video tasks, thanks to their ability to model long-range dependencies and capture spatiotemporal correlations [7, 8].”
Automated HFrEF Diagnosis Using an Optimized TimeSformer Model in Echocardiography
Petmezas, G., Papageorgiou, V. E., Vassilikos, V., Pagourelias, E., Tachmatzidis, D., Tsaklidis, G., Katsaggelos, A. K., & Maglaveras, N. (2025) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-025-01758-2 · Cited by 3
Πετμεζάς, Γ., Papageorgiou, V. Ε., PASSMAN, R. S., Rogers, J. A., Stefanopoulos, L., Katsaggelos, A. K., & Maglaveras, N. (2026) · Journal of Electrocardiology · doi:10.1016/j.jelectrocard.2026.154199 · Cited by 3
ReenactFaces: A Specialized Dataset for Reenactment-Based Deepfake Detection
Vanian, V., Petmezas, G., Konstantoudakis, K., & Zarpalas, D. (2025) · Lecture Notes on Data Engineering and Communications Technologies · doi:10.1007/978-3-031-87784-1_35 · Cited by 2
Vanian, V., Πετμεζάς, Γ., Konstantoudakis, K., & Zarpalas, D. (2025) · ACM International Workshop on Multimedia AI against Disinformation (MAD) · doi:10.1145/3733567.3735569 · Cited by 2
MGAMNET: A Mask-Guided Attention Network for Deepfake Detection
Vanian, V., Πετμεζάς, Γ., Almaloglou, E. E. I., & Zarpalas, D. (2025) · IEEE International Conference on Digital Signal Processing (DSP) · doi:10.1109/dsp65409.2025.11075024 · Cited by 2
Wu, Y., Rocha, B. M., Kaimakamis, E., et al. (2023) · doi:10.2139/ssrn.4395317
Research works printed but not linked: 2
- Sabir, E., Cheng, J., Jaiswal, A., AbdAlmageed, W., Masi, I., Natarajan, P.: Recurrent convolutional strategies for face manipulation detection in videos. In: CVPR Workshops (2019)
- Agarwal, S., Farid, H., Gu, Y., He, M., Nagano, K., Li, H.: Protecting world leaders against deep fakes. In: CVPR Workshops (2019)
Cited by
Deepfake Image Detection Using Transfer Learning TechniquesPreprint
PATEL, H. (2026) · Research Square · doi:10.21203/rs.3.rs-9617329/v1
Explainability-guided local–global fusion for video deepfake detection
Keerthana, S., & Rajam, V. M. A. (2026) · Signal Image and Video Processing · doi:10.1007/s11760-026-05658-y
Li, Q., Li, X., He, X., Huang, H., & Yu, P. (2026) · Flow Measurement and Instrumentation · doi:10.1016/j.flowmeasinst.2026.103456 · Cited by 1
Nirmala., M., & Gowr, P. S. (2026) · International Conference on Intelligent and Sustainable AI Systems (ICOSAAS) · doi:10.1109/icosaas68663.2026.11649324
Hybrid CNN and Transformer-Based Temporal Modeling for Deepfake Video Detection
Kulkarni, A., Patil, R., Yakkundi, S., & Salake, S. V. (2026) · International Conference on Intelligent Technologies (CONIT) · doi:10.1109/conit69683.2026.11621386
Metadata
- Authors
- Petmezas, G.
- ORCID
- Petmezas, Georgios
- Venue
- Machine Vision and Applications
- Published
- 2026-04-07
- Volume
- 37
- Issue
- 3
- ISSN
- 0932-8092 (ISSN-L)
- 1432-1769
- OpenAlex
- W7151598350
- Cited by
- 6
- References
- 55
- Citations by year
- 2026: 6
- FWCI
- 17.15
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- Machine Vision and Applications (publishedVersion, cc-by)
- Funders
- University of Western Macedonia
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Generative Adversarial Networks and Image Synthesis
- Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
- Keywords
- cross-dataset generalization, TimeSformer, transformer models, convolutional neural network architecture, fine-tuning, spatio-temporal modeling, misinformation, temporal context
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-09-29, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-09-29 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.

