Towards Understanding Perceptual Differences between Genuine and Face-Swapped Videos
Abstract
In this paper, we report on perceptual experiments indicating that there are distinct and quantitatively measurable differences in the way we visually perceive genuine versus face-swapped videos.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
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?
Wohler et al.'s study found that the distribution of eye fixations differed between real and high-quality face-swapped videos: for real and low-quality videos, participants tended to fixate on the nose, mouth, and eyes with no significant difference, but for high-quality face-swapped videos, they fixated on the nose and mouth with fewer fixations on the eyes.
“In Wohler et al. ' s [35] study, the distribution of eye fi xations di ff ered between real and high-quality face-swapped videos. For both real and low-quality videos, participants tended to fi xate on the nose, mouth, and eyes, with no signi fi cant di ff erence in their distributions. However, for high-quality face-swapped videos, participants tended to fi xate on the nose and mouth, with fewer”
From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [35]
Higher-quality deepfakes are associated with fewer visual artifacts, making them more difficult to detect.
“Higher-quality deepfakes are associated with fewer visual artifacts, making them more di ffi cult to detect [20, 35, 36, 42].”
From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [35]
Video features such as visual artifacts can impact deepfake detection rates.
“Thus, video features such as these can impact detection rates [2, 20, 22, 27, 35].”
From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [35]
Lower-quality deepfakes are generally easier to detect.
“Second, while lower-quality deepfakes are generally easier to detect [20, 35, 36, 42], imperfections in authentic videos (e.g., lighting inconsistencies or unusual framing) can lead viewers to miscategorize genuine content as manipulated [2, 22, 27, 35, 37].”
From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [35]
Imperfections in authentic videos (e.g., lighting inconsistencies or unusual framing) can lead viewers to miscategorize genuine content as manipulated.
“Second, while lower-quality deepfakes are generally easier to detect [20, 35, 36, 42], imperfections in authentic videos (e.g., lighting inconsistencies or unusual framing) can lead viewers to miscategorize genuine content as manipulated [2, 22, 27, 35, 37].”
From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [35]
Cites
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 1× · see citing works
“Among these, one contains short clips ( 10 seconds ) in a controlled environment [39], one uses publicly available videos of celebrities [45], and others focus on news shows [26, 58, 59].”
BackgroundVaccari, C., & Chadwick, A. (2020) · Social Media + Society · doi:10.1177/2056305120903408 · Cited by 978 · see citing works
DeeperForensics-1.0: A Large-Scale Dataset for Real-World Face Forgery Detection
Jiang, L., Li, R., Wu, W., Qian, C., & Loy, C. C. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00296 · Cited by 599 · see citing works
Do (microtargeted) deepfakes have real effects on political attitudes?
Dobber, T., Metoui, N., Trilling, D., Helberger, N., & de Vreese, C. (2021) · The International Journal of Press/Politics · doi:10.1177/1940161220944364 · Cited by 256 · see citing works
Zimmermann, F., & Kohring, M. (2020) · Political Communication · doi:10.1080/10584609.2019.1686095 · Cited by 227 · cited 1× · see citing works
“One feld with enormous potential to be abused is politics, where face swaps are feared to increase mistrust in politicians and parties, which may drive voters towards more radical groups [18, 63, 73].”
BackgroundIs it an animal? Is it a human face? Fast processing in upright and inverted natural scenes
Rousselet, G. A., Macé, M. J.-M., & Fabre-Thorpe, M. (2003) · Journal of Vision · doi:10.1167/3.6.5 · Cited by 215 · see citing works
Probabilistic learning of emotion categories
Plate, R. C., Wood, A., Woodard, K., & Pollak, S. D. (2019) · Journal of Experimental Psychology: General · doi:10.1037/xge0000529 · Cited by 45 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Deepfake detection by human crowds, machines, and machine-informed crowds
Groh, M., Epstein, Z., Firestone, C., & Picard, R. (2022) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2110013119 · Cited by 248 · see citing works
Gamage, D., Ghasiya, P., Bonagiri, V., Whiting, M. E., & Sasahara, K. (2022) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3491102.3517446 · Cited by 79 · see citing works
Human Performance in Deepfake Detection: A Systematic Review
Somoray, K., Miller, D. J., & Holmes, M. (2025) · Human Behavior and Emerging Technologies · doi:10.1155/hbe2/1833228 · Cited by 19 · see citing works
Sütterlin, S., Ask, T. F., Mägerle, S., Glöckler, S., Wolf, L., Schray, J., Chandi, A., Bursac, T., Khodabakhsh, A., Knox, B. J., et al. (2023) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-35017-7_18 · Cited by 16 · see citing works
- Authors
- Wöhler, L., Zembaty, M., Castillo, S., & Magnor, M.
- Affiliations
- Technische Universität Braunschweig (DE)
- Venue
- Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems
- Publisher
- ACM
- Published
- 2021-05-06
- Pages
- 1-13
- OpenAlex
- W3161544020
- Cited by
- 24
- References
- 7
- Citations by year
- 2026: 3 · 2025: 4 · 2024: 7 · 2023: 4 · 2022: 4 · 2021: 2
- FWCI
- 1.39
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- CHI Conference on Human Factors in Computing Systems (publishedVersion)
- Funders
- Deutsche Forschungsgemeinschaft (MA2555/15-1)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Image and Video Quality Assessment
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
- Keywords
- face swapping, perceptual differences, video authenticity, face manipulation detection, visual perception
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
