Detecting Deepfake Videos with Temporal Dropout 3DCNN
Open accessAbstract
While the abuse of deepfake technology has brought about a serious impact on human society, the detection of deepfake videos is still very challenging due to their highly photorealistic synthesis on each frame. To address that, this paper aims to leverage the possible inconsistent cues among video frames and proposes a Temporal Dropout 3-Dimensional Convolutional Neural Network (TD-3DCNN) to detect deepfake videos. In the approach, the fixed-length frame volumes sampled from a video are fed into a 3-Dimensional Convolutional Neural Network (3DCNN) to extract features across different scales and identified whether they are real or fake. Especially, a temporal dropout operation is introduced to randomly sample frames in each batch. It serves as a simple yet effective data augmentation and can enhance the representation and generalization ability, avoiding model overfitting and improving detecting accuracy. In this way, the resulting video-level classifier is accurate and effective to identify deepfake videos. Extensive experiments on benchmarks including Celeb-DF(v2) and DFDC clearly demonstrate the effectiveness and generalization capacity of our approach.
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
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?
Existing video-level approaches are not specially designed for spatiotemporal inconsistency and cannot properly capture spatiotemporal cues hidden in deepfake videos, making them poorly generalized and vulnerable, even unable to achieve comparable performance to image-level approaches.
“Existing video-level approaches are not specially designed for this inconsistency and can not properly capture the spatiotemporal cues which hide in deepfake videos, which makes these approaches poorly generalized and vulnerable, even cannot achieve comparable performance to imagelevel approaches.”
From Deepfake Video Detection with Spatiotemporal Dropout Transformer (2022) · cites this paper as [50]
Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
Experimental — accuracy not yet measuredThe main reason for the performance is that the approach focuses on capturing patch-level inconsistency from both inter-frame and intra-frames rather than inter-frame inconsistency like [50].
“The main reason is that our approach focuses on capturing the patch-level inconsistency from both inter-frame and intra-frames rather than inter-frame inconsistency like [50] .”
From Deepfake Video Detection with Spatiotemporal Dropout Transformer (2022) · cites this paper as [50]
Remarkable improvement is achieved by our spatiotemporal dropout compared to TD-3DCNN [50] using different dropout operations and network architectures.
“remarkable improvement is achieved by our spatiotemporal dropout.”
From Deepfake Video Detection with Spatiotemporal Dropout Transformer (2022) · cites this paper as [50]
Many previous video-level detectors choose traditional CNN backbones, which may restrict their performance because CNNs have limited representation capacity compared to transformers.
“Many previous video-level detectors choose traditional CNN as their backbones, which may restrict their performance since CNN's limited representation capacity compared to transformer, such as [50].”
From Deepfake Video Detection with Spatiotemporal Dropout Transformer (2022) · cites this paper as [50]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 38 · Cited by 77
Research works printed but not linked: 4
- [7]7. Chu et al., 2020] Mengyu Chu, You Xie, Jonas Mayer, Laura Leal- Taixé, and Nils Thuerey. Learning temporal coherence via self-supervision for gan-based video generation. ACM TOG, 39(4):75:1-13, 2020.
- [17]17. Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron C. Courville, and Yoshua Bengio. Generative adversar- ial nets. In NeurIPS, pages 2672-2680, 2014.
- [33]33. F. Matern, C. Riess, and M. Stamminger. Ex- ploiting visual artifacts to expose deepfakes and face manipula- tions. In WACVW, pages 83-92, 2019.
- [49]49. Rubén Tolosana, Sergio Romero-Tapiador, Julian Fiérrez, and Rubén Vera-Rodríguez. Deepfakes evolution: Analysis of facial regions and fake detection performance. In ICPR Workshops(5), pages 442-456, 2020.
- Authors
- Zhang, D., Li, C., Lin, F., Zeng, D., & Ge, S.
- Affiliations
- Chinese Academy of Sciences (CN)
- Shanghai University (CN)
- Venue
- Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence
- Published
- 2021-08
- Pages
- 1288-1294
- OpenAlex
- W3188946793
- Semantic Scholar
- 8b65ccd6e2b878639cc4bd01f6f0aea8d22e60b2
- Cited by
- 77
- References
- 38
- Citations by year
- 2026: 5 · 2025: 18 · 2024: 27 · 2023: 16 · 2022: 10 · 2021: 1
- FWCI
- 4.42
- Type
- conference paper
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- International Joint Conference on Artificial Intelligence (publishedVersion)
- Funders
- National Natural Science Foundation of China (61772513)
- Chinese Academy of Sciences
- Youth Innovation Promotion Association of the Chinese Academy of Sciences
- Beijing Municipal Science and Technology Commission (Z191100007119002)
- Youth Innovation Promotion Association
- 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
- deepfake detection, 3D convolutional neural networks, data augmentation, overfitting, Celeb-DF, DFDC
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
