A Dual-Branch Fusion Model for Deepfake Detection Using Video Frames and Microexpression Features
Open accessThe DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.
Abstract
Deepfake detection has become a critical issue due to the rise of synthetic media and its potential for misuse. In this paper, we propose a novel approach to deepfake detection by combining video frame analysis with facial microexpression features. The dual-branch fusion model utilizes a 3D ResNet18 for spatiotemporal feature extraction and a transformer model to capture microexpression patterns, which are difficult to replicate in manipulated content. We evaluate the model on the widely used FaceForensics++ (FF++) dataset and demonstrate that our approach outperforms existing state-of-the-art methods, achieving 99.81% accuracy and a perfect ROC-AUC score of 100%. The proposed method highlights the importance of integrating diverse data sources for deepfake detection, addressing some of the current limitations of existing systems.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures



Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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
“The widespread availability of generative AI has increased the volume of images and videos produced and circulated through online platforms [1][2][3].”
From A Study on Broker-Assisted Blockchain Trust Chains for Provenance and Integrity Verification of Generative Media Using Watermarking, Semantic Fingerprinting, and C2PA (2026) · cites this paper as [1]
Experimental — accuracy not yet measured
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
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404 · cited 1×
“On the other hand, transformer [25] models, originally introduced for natural language processing, have gained prominence in vision tasks due to their ability to model long-range dependencies and relationships.”
Ekman, P., & Friesen, W. V. (2019) · PsycTESTS Dataset · doi:10.1037/t27734-000 · Cited by 6,019
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 · cited 1×
“For our experiments, we utilized two widely recognized deepfake detection benchmarks, namely FF++ [16] and Celeb-DF [17], as well as a recently introduced deepfake dataset named ReenactFaces [18].”
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×
“For our experiments, we utilized two widely recognized deepfake detection benchmarks, namely FF++ [16] and Celeb-DF [17], as well as a recently introduced deepfake dataset named ReenactFaces [18].”
Multi-attentional Deepfake Detection
Zhao, H., Wei, T., Zhou, W., Zhang, W., Chen, D., & Yu, N. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00222 · Cited by 1,019 · cited 2×
“…Zhang et al. [5] 95.20% 98.68% Miao et al. [6] 98.71% 99.74% Zhang et al. [7] 94.14% 98.44% Guo et al. [8] 96.01% 98.92% Long et al. [9] 97.64% 99.70% Pang et al. [10] 97.76% 98.81% Yang et al. [11] 97.86% 99.38% Zhao et al. [12] 97.60% 99.29% Alfalasi et al. [15] 97.31% 99.60% Ours 99.81% 100%”
InfluentialLearning Spatio-Temporal Features with 3D Residual Networks for Action Recognition
Hara, K., Kataoka, H., & Satoh, Y. (2017) · 2017 IEEE International Conference on Computer Vision Workshops (ICCVW) · doi:10.1109/iccvw.2017.373 · Cited by 726 · cited 1×
“The 3D ResNet18 [24] extends the ResNet architecture to video data by incorporating 3D convolutions, enabling it to learn spatiotemporal features from sequences of frames.”
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 · cited 1×
“The emergence of deepfake technology has created serious concerns about the trust-worthiness of digital media [1].”
Cross-dataset learning and person-specific normalisation for automatic Action Unit detection
Baltrusaitis, T., Mahmoud, M., & Robinson, P. (2015) · 2015 11th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG) · doi:10.1109/fg.2015.7284869 · Cited by 412 · cited 1×
“To leverage this, we followed the methodology by Baltrušaitis et al. [22] to perform a feature-based analysis of facial landmarks and AUs, extracting the relevant microexpression features from the preprocessed video frames.”
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×
“In response to the growing threat of deepfakes, researchers have developed automated detection methods that primarily analyze visual and temporal inconsistencies in video frames [2].”
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 · cited 1×
“We further benchmarked our method on ReenactFaces, a new dataset comprising 10,000 real and 10,000 re-enacted videos created using three state-of-the-art facial re-enactment methods, namely the First-Order Motion Model [19], Thin-Plate Spline Motion [20], and FSRT [21].”
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 · cited 1×
“We further benchmarked our method on ReenactFaces, a new dataset comprising 10,000 real and 10,000 re-enacted videos created using three state-of-the-art facial re-enactment methods, namely the First-Order Motion Model [19], Thin-Plate Spline Motion [20], and FSRT [21].”
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 · cited 1×
“However, as deepfake technology continues to improve, generating increasingly realistic facial appearances and expressions, existing approaches struggle to detect subtle manipulations [3].”
MTD-Net: Learning to Detect Deepfakes Images by Multi-Scale Texture Difference
Yang, J., Li, A., Xiao, S., Lu, W., & Gao, X. (2021) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2021.3102487 · Cited by 184 · cited 3×
“…Zhang et al. [5] 95.20% 98.68% Miao et al. [6] 98.71% 99.74% Zhang et al. [7] 94.14% 98.44% Guo et al. [8] 96.01% 98.92% Long et al. [9] 97.64% 99.70% Pang et al. [10] 97.76% 98.81% Yang et al. [11] 97.86% 99.38% Zhao et al. [12] 97.60% 99.29% Alfalasi et al. [15] 97.31% 99.60% Ours 99.81% 100%”
InfluentialF 2 Trans: High-Frequency Fine-Grained Transformer for Face Forgery Detection
Miao, C., Tan, Z., Chu, Q., Liu, H., Hu, H., & Yu, N. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2022.3233774 · Cited by 172 · cited 3×
“Similarly, Miao et al. [6] proposed F2Trans, which combines wavelet sampling and central difference attention for frequency-aware feature extraction.”
InfluentialAn Overview of Facial Micro-Expression Analysis: Data, Methodology and Challenge
Xie, H., Lo, L., Shuai, H.-H., & Cheng, W.-H. (2022) · IEEE Transactions on Affective Computing · doi:10.1109/taffc.2022.3143100 · Cited by 98 · cited 1×
“Microexpressions are innate, brief facial movements that occur in response to emotions and typically last for less than a second [4].”
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
Self-citationMRE-Net: Multi-Rate Excitation Network for Deepfake Video Detection
Pang, G., Zhang, B., Teng, Z., Qi, Z., & Fan, J. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3239607 · Cited by 72 · cited 2×
“Pang et al. [10] captured spatiotemporal dependencies through a multi-rate excitation network.”
InfluentialFace Forgery Detection via Multi-Feature Fusion and Local Enhancement
Zhang, D., Chen, J., Liao, X., Li, F., Chen, J., & Yang, G. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3390945 · Cited by 62 · cited 2×
“Zhang et al. [7] used dual-domain fusion and local enhancement attention to enrich feature representation.”
InfluentialGuo, Z., Wang, L., Yang, W., Yang, G., & Li, K. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3289147 · Cited by 60 · cited 2×
“Guo et al. [8] introduced LDFNet, a lightweight detector that fuses local visual artifacts and global texture cues via a dynamic fusion module and TraceBlock-based inference.”
InfluentialWATCHER: Wavelet-Guided Texture-Content Hierarchical Relation Learning for Deepfake Detection
Wang, Y., Chen, C., Zhang, N., & Hu, X. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02116-5 · Cited by 44 · cited 1×
“Meanwhile, Wang et al. [14] introduced WATCHER, which employs wavelet-guided hierarchical learning to jointly reason over texture and content features through multi-domain fusion and attention mechanisms, achieving superior generalization to unseen manipulations.”
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 · cited 1×
“We further benchmarked our method on ReenactFaces, a new dataset comprising 10,000 real and 10,000 re-enacted videos created using three state-of-the-art facial re-enactment methods, namely the First-Order Motion Model [19], Thin-Plate Spline Motion [20], and FSRT [21].”
LGDF-Net: Local and Global Feature-Based Dual-Branch Fusion Networks for Deepfake Detection
Long, M., Liu, Z., Zhang, L., & Peng, F. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3530402 · Cited by 22 · cited 3×
“Long et al. [9] proposed LGDF-Net, which explicitly separates and processes local and global features using specialized modules and fuses them through a multi-level strategy.”
InfluentialFace Forgery Detection Based on Fine-Grained Clues and Noise Inconsistency
Zhang, D., He, R., Liao, X., Li, F., Chen, J., & Yang, G. (2024) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2024.3455311 · Cited by 19 · cited 2×
“For instance, Zhang et al. [5] introduced a two-stream architecture with a multiscale transformer module and a fusion mechanism to analyze spatial and noise flow artifacts.”
InfluentialMining Temporal Inconsistency with 3D Face Model for Deepfake Video Detection
Cheng, Z., Chen, C., Zhou, Y., & Hu, X. (2023) · Lecture notes in computer science · doi:10.1007/978-981-99-8540-1_19 · Cited by 11 · cited 1×
“In a related direction, Cheng et al. [13] proposed leveraging 3D facial geometry by measuring inconsistencies between 2D landmarks and 3D reprojections across video frames, capturing subtle temporal anomalies using an RNN-based framework.”
Generalizable Spatiotemporal Deepfake Detection with Late-Fusion
Alfalasi, H. R., Hashem, M., & Abul, O. (2024) · International Conference on Signal Processing and Information Security (ICSPIS) · doi:10.1109/icspis63676.2024.10812603 · Cited by 2 · cited 2×
“…Zhang et al. [5] 95.20% 98.68% Miao et al. [6] 98.71% 99.74% Zhang et al. [7] 94.14% 98.44% Guo et al. [8] 96.01% 98.92% Long et al. [9] 97.64% 99.70% Pang et al. [10] 97.76% 98.81% Yang et al. [11] 97.86% 99.38% Zhao et al. [12] 97.60% 99.29% Alfalasi et al. [15] 97.31% 99.60% Ours 99.81% 100%”
InfluentialReenactFaces: 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 · cited 2×
“For our experiments, we utilized two widely recognized deepfake detection benchmarks, namely FF++ [16] and Celeb-DF [17], as well as a recently introduced deepfake dataset named ReenactFaces [18].”
Self-citation
Research works printed but not linked: 1
- Kay, W., Carreira, J., Simonyan, K., Zhang, B., Hillier, C., Vijayanarasimhan, S., Viola, F., Green, T., Back, T., and Natsev, A. (2017). The Kinetics Human Action Video Dataset. arXiv.
Cited by
Cross-dataset video deepfake detection using Transformer and CNN architectures
Petmezas, G. (2026) · Machine Vision and Applications · doi:10.1007/s00138-026-01809-w · Cited by 6
“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].”
Yang, C., & Kim, M. (2026) · Applied Sciences · doi:10.3390/app16073391
Technical Framework for Combating Deepfake Identity Fraud in E-Government Services
Hanna, E., Zahr, E. A., Samrouth, K., & Bakir, N. (2026) · International Symposium on Digital Forensics and Security (ISDFS) · doi:10.1109/isdfs69419.2026.11458925
“3) Micro-expression and head-pose analysis: A method that uses a combination of both frame analysis and micro-expression features by using a dual branch model [13].”
Blink-and-You’ll-Miss-It: Micro-expressions as a Deepfake Detection Add-On
Brad, A., & Pungilă, C. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-29251-3_1
Deepfake detection: A review of audio, visual and multimodal techniques
Kanwal, R., Ashraf, H., Jhanjhi, N. Z., & Mirza, Q. U. A. Z. Z. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.115845
Metadata
- Authors
- Petmezas, G., Vanian, V., Rufete, M. P., Almaloglou, E. E. I., & Zarpalas, D.
- Affiliations
- Centre for Research and Technology Hellas, 57001 Thessaloniki, Greece
- Centre for Research and Technology Hellas (GR)
- Herta Security, 08037 Barcelona, Spain
- Corresponding author
- Vanian, Vazgken
- Venue
- Journal of Imaging
- Publisher
- MDPI AG
- Published
- 2025-07-11
- Volume
- 11
- Issue
- 7
- Pages
- 231
- ISSN
- 2313-433X (ISSN-L)
- PMC
- PMC12295270
- OpenAlex
- W4412196232
- Cited by
- 8
- References
- 26
- Citations by year
- 2026: 8
- FWCI
- 2.81
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Journal of Imaging (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- European Commission (101073928)
- 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 › Computer Vision and Pattern Recognition › Generative Adversarial Networks and Image Synthesis
- Keywords
- deepfake detection, facial micro-expressions, video frame analysis, spatiotemporal feature extraction, transformer models, synthetic media, FaceForensics++
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
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.