The DOI matches its Crossref record; 38 of the 39 references it cites were found in a registry or an index.
Detecting Deep-Fake Videos from Appearance and Behavior
Open accessAbstract
Synthetically-generated audios and videos - so-called deep fakes - continue to capture the imagination of the computer-graphics and computer-vision communities. At the same time, the democratization of access to technology that can create a sophisticated manipulated video of anybody saying anything continues to be of concern because of its power to disrupt democratic elections, commit small to large-scale fraud, fuel disinformation campaigns, and create non-consensual pornography. We describe a biometric-based forensic technique for detecting face-swap deep fakes. This technique combines a static biometric based on facial recognition with a temporal, behavioral biometric based on facial expressions and head movements, where the behavioral embedding is learned using a CNN with a metric-learning objective function. We show the efficacy of this approach across several large-scale video datasets, as well as in-the-wild deep fakes.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Existing unimodal deepfake detectors are primarily designed to detect a single type of manipulation, such as visual, acoustic, and text.”
From Contextual Cross-Modal Attention for Audio-Visual Deepfake Detection and Localization (2024) · cites this paper as [2]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“the re- cent common underlying mechanism for deepfake cre- ation is deep learning models such as autoencoders and generative adversarial networks (GANs)”
From Deep learning for deepfakes creation and detection: A survey (2022) · cites this paper as [2]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
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 · cited 1×
“We used the ResNet-101 network architecture [24], where the input layer of the network is modified to the size of our feature matrix (256×100).”
MethodGAN（Generative Adversarial Nets）
淳司, 柴. (2017) · Journal of Japan Society for Fuzzy Theory and Intelligent Informatics · doi:10.3156/jsoft.29.5_177_2 · Cited by 20,507
A Style-Based Generator Architecture for Generative Adversarial Networks
Karras, T., Laine, S., & Aila, T. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00453 · Cited by 11,714 · cited 1×
“In the static image domain, highly realistic images of people can now be synthesized using generative adversarial networks (GANs) [2], [3].”
BackgroundTowards Evaluating the Robustness of Neural Networks
Carlini, N., & Wagner, D. (2017) · 2017 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2017.49 · Cited by 8,859 · cited 1×
“The drawback is that they can be highly sensitive to intentional or unintentional laundering including resizing or trans-coding, as well as adversarial attacks [7] and extrapolation to novel datasets.”
BackgroundAnalyzing and Improving the Image Quality of StyleGAN
Karras, T., Laine, S., Aittala, M., Hellsten, J., Lehtinen, J., & Aila, T. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00813 · Cited by 6,332
Estimating the Support of a High-Dimensional Distribution
Schölkopf, B., Platt, J. C., Shawe-Taylor, J., Smola, A. J., & Williamson, R. C. (2001) · Neural Computation · doi:10.1162/089976601750264965 · Cited by 6,168
Parkhi, O. M., Vedaldi, A., & Zisserman, A. (2015) · Procedings of the British Machine Vision Conference 2015 · doi:10.5244/c.29.41 · Cited by 5,479 · cited 11×
“VGG yields a distinct 4096-”
MethodInfluentialWaveNet: a generative model for raw audio
Oord, A. V. D., Dieleman, S., Zen, H., Simonyan, K., Vinyals, O., Graves, A., Kalchbrenner, N., Senior, A., & Kavukcuoglu, K. (2016) · arXiv · doi:10.48550/arxiv.1609.03499 · Cited by 4,911 · cited 1×
“In the audio domain, highly realistic audio synthesis is now possible in which a neural network, with enough sample recordings, can learn to synthesize speech in your voice [25].”
BackgroundFaceForensics++: 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 · cited 1×
“The FaceForensics++ dataset (FF) [27] consists of 1000 YouTube videos of 1000 different people, mostly news anchors”
VoxCeleb2: Deep Speaker Recognition
Chung, J. S., Nagrani, A., & Zisserman, A. (2018) · Interspeech 2018 · doi:10.21437/interspeech.2018-1929 · Cited by 2,256
Cited by
Sorted by citations
Items 201–245 of 245
Watolla, A.-K. M. (2024) · Digitale Gesellschaft · doi:10.14361/9783839473115-001
Watolla, A.-K. M. (2024) · Digitale Gesellschaft · doi:10.14361/9783839473115-fm
Unmasking Deepfakes: Advances in Fake Video Detection
Odeh, A. H. (2024) · Revue d intelligence artificielle · doi:10.18280/ria.380407 · Cited by 1
RuleBoost: A Neuro-Symbolic Framework for Robust Deepfake Detection
Raza, M. A., Malik, K. M., & Haq, I. U. (2024) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb62174.2024.10744498
Güncel Deepfake Video Algılama Yöntemleri Üzerine Bir Literatür İncelemesi
Aydın, S., & Özen, Z. (2024) · Bilgisayar bilimleri ve mühendisliği dergisi. · doi:10.54525/bbmd.1460699
Explainable Recognition System for Computer-Modified Face Photos
Kamiński, A., Degórski, K., Kaźmierczak, S., & Mańdziuk, J. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-84356-3_23
PerceptiGuard: Unveiling DeepFakes Through Enhanced Media Perception
Phatangare, S., Tekade, O., Thakare, P. P., Wankhade, A., & Wankhade, P. (2024) · IEEE Silchar Subsection Conference (SILCON) · doi:10.1109/silcon63976.2024.10910356
“In [9], the authors have explored advanced techniques in detecting deepfake videos using deep learning methodologies.”
A., M., Pushpitha Shree, P., & Brindha, G. (2024) · International Conference on IoT, Communication and Automation Technology (ICICAT) · doi:10.1109/icicat62666.2024.10923429
“Agarwal et al. [11] employed temporal and static biometrics based on head movements and facial expressions to present a biometric forensic approach for identifying face-swap deep fakes using a CNN with a metric learning objective function.”
Sheikh, M. F. A., Bhuyan, A. S., Dhiman, A., Sharma, C., Diya, Joshi, M., & Ritu, R. (2024) · Asian Conference on Intelligent Technologies (ACOIT) · doi:10.1109/acoit62457.2024.10941577
“They stumble on irregularities in face traits and motions via the use of device learning, which provides a feasible route for reinforcing detection accuracy[4].”
Manisha, A., Sri, S. R., & Vinya, V. L. (2024) · International Conference on Artificial Intelligence Trends and Pattern Recognition (ICAITPR) · doi:10.1109/icaitpr63242.2024.10960037
A Novel Unified Approach to Deepfake Detection of Images
Sen, L., & Mukherjee, S. (2024) · International Conference on Intelligent Systems and Embedded Design (ISED) · doi:10.1109/ised63599.2024.10957205 · Cited by 2
“An architecture to detect deepfakes by CNN is presented in [3], [7], audio analysis in [8], cross domain analysis is done in [1], [10] to [16], multi-attentional approach in [6], vission transformer in [9], using Fourier transform in [5].”
Advances in Deepfake Detection: Technologies, Methods and Future Directions
Tejaswi, S., Babu, T., & Tejaswi, K. (2025) · International Conference on Data Science and Business Systems (ICDSBS) · doi:10.1109/icdsbs63635.2025.11032017
Improving Image Captioning Accuracy Using Advanced Deep Learning Techniques
Chandar, J. N., Kavitha, G., & Sethuraman, A. (2025) · Annals of Emerging Technologies in Computing · doi:10.33166/aetic.2025.02.004
DIDM: Deepfake Image Detection Model Using Deep Learning
Misra, T., & Dinker, A. G. (2025) · International Conference on Networks and Cryptology (NETCRYPT) · doi:10.1109/netcrypt65877.2025.11102670
“[5] For example, in cybersecurity, deepfakes have been used for identity theft and social engineering attacks, highlighting the need for reliable detection mechanisms.”
Detecting Forgery in War Footage for Information Accuracy
Afzal, S., Sharafudeen, M., & S, V. C. S. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-4008-9_38
Sohrawardi, S. J., Wu, Y.-C., & Wright, M. (2025) · Springer eBooks · doi:10.1007/978-3-031-89853-2_10
Seeing is No Longer Believing: How Deepfakes May Shape the Future of Identity Credibility in Media
Frick, R. A., Bunzel, N., & Graner, L. (2025) · ACM Workshop on Deepfake, Deception, and Disinformation Security (3D-Sec) · doi:10.1145/3733813.3764371
“For video face swaps, model-based methods analyze behavioral, signal-based, or geometric features [3, 40, 43].”
Computational Analysis of the KnowBe4 Identity Fraud Case by a Multi-Adaptive Dynamical System Model
Boddie, K. L., Brown, E., Hartman, A., Sillie, X., Bouma, D., Jan Treur, A., & Roelofsma, P. H. M. P. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-032-07992-3_42
Multimodal Consistency-Driven Deepfake Detection
Zhang, L., Liu, B., Chu, Q., & Yu, N. (2025) · Lecture notes in computer science · doi:10.1007/978-981-95-3393-0_24
AIGC Video Detection Based on Missing Motion Details
Cao, Y., Tian, Y., Huang, L., Bao, Y., & Li, Q. (2025) · Lecture notes in computer science · doi:10.1007/978-981-95-3729-7_5
Combating Falsification of Speech Videos with Live Optical Signatures
Schwartz, H., Yan, X., Carver, C. J., & Zhou, X. (2025) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3719027.3765112 · Cited by 1
Detecting Deepfakes using Temporal Consistency of Facial Expression Transitions
Chacko, R. E., & Bajwa, G. (2025) · Conference on Privacy, Security and Trust · doi:10.1109/pst65910.2025.11268852
Deep Learning Classification Techniques for Video Tampering Detection: A Review
Khalsa, G. K., & Ahuja, R. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-02790-0_15
A Comprehensive Review on Popular CNN-Based Deepfake Detection Models, Performances, and Challenges
Paul, S., Rangababu, P., & Devi, B. P. (2026) · Lecture notes in networks and systems · doi:10.1007/978-981-96-9932-2_18
TrueSight: Deep Fake Detection Using Xception and Temporal Convolutional Networks
Paul, S., Nath, I., Kundu, S., Datta, S., & Biswas, S. (2026) · Lecture notes in electrical engineering · doi:10.1007/978-981-96-9720-5_2
Teaching People to Spot Deepfakes: Investigating User Acceptance of LLM-Powered Chatbots
Chen, C., Goh, D. H., & Tang, Z. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2622583 · Cited by 1
“This need is particularly pressing for deepfakes, which are multimodal and more difficult to evaluate than other forms of disinformation (Agarwal et al., 2020; Cooke et al., 2025; Groh et al., 2022).”
Spot the Fake: Vision Transformers in the Fight Against DeepFakes
Pandya, K., & DAHIYA, V. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-17837-4_24
Benford-Constrained Multimodal Learning for Few-Shot Fake Image Detection
Astraikis, J., Martin, R., van Es, A., Kottam, S., & Sun, Y. (2025) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata66926.2025.11401004
“Wang et al. [14] developed a unified forensic framework demonstrating improved cross-dataset generalization, while Agarwal and Farid [15] extended detection to behavioral analysis beyond visual artifacts.”
Deepfake Video Detection Framework Employing Deep Learning and Explainability Models
Manivannan, D., Kishore Kumar, K., Durai, S., K, M. N., Vulichi, L., & Mohammed, S. M. (2026) · International Conference on Electronics, Materials Engineering & Nano-Technology (IEMENTech) · doi:10.1109/iementech202669403.2026.11434388
DeepShield Lite-Multimodal Deepfake Detection for the Next Web Era
Sivakumar, V., G, S. M., Sivaprasath, K., Vignesh, G., & Mir, S. (2026) · International Conference on Image Processing and Capsule Networks (ICIPCN) · doi:10.1109/icipcn67432.2026.11438484
“Agarwal, Farid, El-Gaaly, and Lim examined deepfake video detection using a combination of appearance-based and behavioral cues, demonstrating that integrating facial texture analysis with motion and behavioral patterns improves robustness against advanced facial manipulation techniques [15].”
InfluentialDhivya, K., P, N. K. M., R, J. G. G., & Naveen Kumar, B. (2025) · International Conference on Advancement in Futuristic Technologies (ICAFT) · doi:10.1109/icaft66710.2025.11452748
“Although models like ResNet and EfficientNet showed strong detection abilities, they often require significant computing power, large amounts of training data, and struggle to generalize across various manipulation methods and compression artifacts [5], [6].”
Deep Fake Detection and Prevention: A Digital Forensics Perspective
M, P. B., Shivashankara, S., R, P. D., & Sudeep, J. (2025) · International Conference on Emerging Technologies in Electronics and Green Energy (ICETEG) · doi:10.1109/iceteg66194.2025.11472840
“Agarwal et al. [15] demonstrated that combining lip movement analysis with audio cues improves robustness under compression.”
Video Deepfake Detection: Challenges and Recent Trends
Koutlis, C., Pianese, A., Cozzolino, D., Schinas, M., Mylonas, S., Verdoliva, L., & Papadopoulos, S. (2026) · Springer eBooks · doi:10.1007/978-3-032-11782-3_8
STRIKE: A Structured Taxonomy of Cybercrime for Risk, Impact, Knowledge, and Evolution
Pappy, M., Nguyen, L., Kumar, S., Jung, B., & Chen, B. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-14893-3_14
“Deep fake can be active or passive, active deep fake includes Face and voice swapping in audio and videos to impersonate a person in live video calls [12], causing business-related authentication to be broken than relying on audio and video of a person.”
A unified hybrid approach to deepfake video recognition
Shaikh, A., Mittal, S., Kumari, R., & Jindal, S. (2026) · AIP conference proceedings · doi:10.1063/5.0330163
Digital Deception: A Comparative Analysis of CNN and CNN-LSTM Models for Deepfake Detection
Kamthania, D., Palia, N., & Chopra, D. (2026) · International Conference on Computing for Sustainable Global Development · doi:10.23919/indiacom70271.2026.11525770
“Agarwal et al. presented a biometric-based forensic technique for detecting face-swap deepfakes by combining temporal behavioural biometrics acquired by a CNN with static facial recognition for real world scenarios and large-scale datasets [8].”
Forensics Adapter: Unleashing CLIP for Generalizable Face Forgery Detection
Cui, X., Li, Y., Zhu, D., Zhou, J., Dong, J., & Lyu, S. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3706492 · Cited by 9
“…capture the forgery traces, utilizing features ranging from the biological signals (e.g., eye blinking [33], heartbeat rhythm [15], head pose [14], facial action units [35]), frequency signals (e.g., [19], [20], [21]), and auto-learned clues with dedicated training schemes (e.g., [24], [36], [37]).”
Iqbal, S., & Munir, M. W. (2026) · International Conference on IT and Industrial Technologies (ICIT) · doi:10.1109/icit68548.2026.11577683
Spatio-Temporal Approach to Deepfake Detection Using Periocular Physiological Cues
Mohod, N., Sable, A. V., Balapure, S., Avatirak, R., Tayde, A., Sahu, N., & Girnare, P. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-27120-4_17
Deep-fake detection of digital misinformation over social media: a comprehensive review
Abdelsalam, H. H. A., & Hefny, H. A. (2026) · Array · doi:10.1016/j.array.2026.101086
A Real-Time Face Swap Detection Model for Video Chatting Scams
Li, X., Weng, Y., Zhang, J., Liu, H., & Wang, H. (2026) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-032-22542-9_20
Digital Manipulation of Democracy:
Rodríguez, R., Kolloju, N., Chaketi, R., & Boddu, V. (2026) · Dia-noesis. · doi:10.12681/dia.46978
Shiohara, K., Yamasaki, T., & Golyanik, V. (2026) · doi:10.1109/cvprf71909.2026.00355
“It is noteworthy that, differently from the previous reference-assisted approaches [2, 17, 18] that perform large-scale pretraining on identity-annotated datasets such as VoxCeleb2 [14], our model is pre-trained only on unlabeled videos, which enables further extension on larger video collections.”
Metadata
- Authors
- Agarwal, S., Farid, H., El-Gaaly, T., & Lim, S.-N.
- Affiliations
- University of California, Berkeley (US)
- Meta (United States) (US)
- Venue
- 2020 IEEE International Workshop on Information Forensics and Security (WIFS)
- Publisher
- IEEE
- Published
- 2020-12-06
- Pages
- 1-6
- ISSN
- 2157-4774 (ISSN-L)
- OpenAlex
- W3021589373
- Cited by
- 242
- References
- 39
- Citations by year
- 2026: 26 · 2025: 39 · 2024: 79 · 2023: 36 · 2022: 36 · 2021: 22 · 2020: 4
- FWCI
- 7.56
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE International Workshop on Information Forensics and Security (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Advanced Research Projects Agency
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
- Physical Sciences › Computer Science › Signal Processing › Biometric Identification and Security
- Keywords
- deepfake detection, face recognition, facial expressions, head movement, behavioral biometrics, metric learning, disinformation
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.
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
