Multisemantic Path Neural Network for Deepfake Detection
Open accessThe DOI matches its Crossref record; 61 of the 61 references it cites were found in a registry or an index; 1 of the 15 sources its claims cite were checked for support.
Abstract
With the continuous development of deep learning techniques, it is now easy for anyone to swap faces in videos. Researchers find that the abuse of these techniques threatens cyberspace security; thus, face forgery detection is a popular research topic. However, current detection methods do not fully use the semantic features of deepfake videos. Most previous work has only divided the semantic features, the importance of which may be unequal, by experimental experience. To solve this problem, we propose a new framework, which is the multisemantic pathway network (MSPNN) for fake face detection. This method comprehensively captures forged information from the dimensions of microscopic, mesoscopic, and macroscopic features. These three kinds of semantic information are given learnable weights. The artifacts of deepfake images are more difficult to observe in a compressed video. Therefore, preprocessing is proposed to detect low-quality deepfake videos, including multiscale detail enhancement and channel information screening based on the compression principle. Center loss and cross-entropy loss are combined to further reduce intraclass spacing. Experimental results show that MSPNN is superior to contrast methods, especially low-quality deepfake video detection.
Overview
CoChat Summary
TL;DR
The paper proposes MSPNN, a multisemantic path neural network that fuses microscopic, mesoscopic, and macroscopic features with learnable channel-spatial attention weights for deepfake detection. It reports strong accuracy on FaceForensics++ and DeepFake-TIMIT, especially for low-quality/high-compression videos, with cross-dataset generalization on Celeb-DFv2.
Key findings
- MSPNN achieved 94.21% accuracy on FF++(HQ/C23) and 76.31% on FF++(LQ/C40), and AUCs of 99.12 (TIMIT HQ) and 99.52 (TIMIT LQ), outperforming the compared methods in Table 5.
- Cross-dataset validation on Celeb-DFv2 gave an AUC of 66.7, while the same model scored 96.7 on FF-DF; the paper notes Masi’s method generalizes better on Celeb-DFv2 (73.41) but has a much lower original-dataset AUC.
- Ablations show preprocessing and attention help some subsets, especially high-compression data: Y-channel input raised FF++(LQ) NeuralTextures from 74.41% (RGB) to 74.92%, and attention raised FF++(LQ) DeepFakes from 93.83% to 94.87% (Tables 2 and 4).
- The three semantic pathways are intended to capture different forgery cues: microscopic details, mesoscopic blocks, and macroscopic regions such as eyes, nose, and mouth, with weights assigned automatically rather than manually.
Methods
- Design: MSPNN uses three semantic pathways based on receptive fields—micro (3×3), meso (two 3×3 replacing 5×5), and macro (three 3×3 replacing 7×7)—plus multiscale detail enhancement with three Gaussian filters (5×5, 9×9, 19×19; σ = 1.0, 2.0, 4.0; weights 0.5, 0.5, 0.25) and YCrCb channel separation using the Y channel. Features are fused via channel-spatial attention, and training combines cross-entropy loss with center loss (λ = 1).
- Data: FaceForensics++ (1,000 YouTube videos; 1,000 fake and 1,000 real videos per compression rate; splits 720 train/140 validation/140 test), DeepFake-TIMIT (640 fake videos at 128×128 and 64×64; splits 230/45/45), and Celeb-DFv2 (590 real, 5,639 deepfake videos) for cross-dataset validation. Faces were extracted with dlib, expanded by 1.3×, and 30 frames per video were taken.
- Implementation/analysis: PyTorch on an RTX 3090; Adam with initial learning rate 1e-4 and cosine annealing, center loss optimized with SGD, batch size 32, 100 epochs, input 224×224, random seed 43. Evaluation used accuracy and AUC, with preprocessing, pathway, attention, and cross-dataset ablations.
Limitations
- The paper states that balancing accuracy and generalization remains a challenge; Masi’s method generalizes better on Celeb-DFv2 than MSPNN, though its original-dataset AUC is far behind.
- The paper states that only intraframe compression was used because removing adjacent frames would conflict with benchmark dataset creation requirements; dlib sometimes failed to recognize faces, and those frames were skipped.
- Evident from the tables: gains from detail enhancement and attention are mixed, with some subsets decreasing (e.g., detail enhancement lowered FF++ HQ DeepFakes from 99.73% to 99.54%, and attention lowered FF++ HQ DeepFakes from 99.73% to 99.49%), so improvements are not uniform across all datasets or compression levels.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
1 Valid · 3 Not valid · 11 Not checked · 46 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Besides, methods based on hard-to-find biological signals, such as heart rate, would be far less accurate due to video compression and other processing [39].”
§ Multisemantic Path Neural Network for Deepfake Detection > 2.2. Deepfake Image Detection.
What the sources say
- Partly supportedcites [39] X. Jin, D. Ye, and C. Chen, "Countering spoof: towards detecting deepfake with multidimensional biological signals," Security and Communication Networks, vol. 2021, no. 1, , Article ID 6626974, 8 pages, 2021. sourceExperimental — accuracy not yet measured
Section 3.1 says heart-rate signals are difficult to extract in unrestricted settings and that denoising and filtering can destroy abnormal signals in fake videos, weakening classification. It does not identify video compression as a cause or establish that biological-signal methods generally are far less accurate; the Introduction reports high accuracy for the authors’ method. Page numbers are unavailable in our copy. (LLM, conf=0.85)
hedgedstates 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
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Experimental results show that MSPNN is superior to contrast methods, especially low-quality deepfake video detection.”
- Also stated in: § 1. Introduction · plain
“Experimental results show that our network has better accuracy than current methods, especially in low-quality deepfake video encoding.”
§ Multisemantic Path Neural Network for Deepfake Detection > Abstract
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.
- Also stated in: § 1. Introduction · plain
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“After comparative experimental analysis, we find that when the Y channel of the image is used as the input, the accuracy is higher than that when other channels are used.”
§ Multisemantic Path Neural Network for Deepfake Detection > 1. Introduction
plainstates an association- 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.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“e proposed preprocessing method can improve the detection of low-quality counterfeit videos, and the network can comprehensively capture different semantic information of images.”
§ Multisemantic Path Neural Network for Deepfake Detection > 1. Introduction
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: unclear — It is unclear why the paper says it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Our addition of preprocessing makes up for this problem, as shown in Table 1.”
§ Multisemantic Path Neural Network for Deepfake Detection > 4.4. Experimental Results.
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.
25 more claims
Create a free account to see all 61 claims and to verify the 14 references still open.
- It may state two separate claims as one. 19
- Probably not a substantial claim. 13
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?
In the citing paper
“| | | | AUC = 99.12 | TIMIT(HQ) |”
From Deepfake Attacks: Generation, Detection, Datasets, Challenges, and Research Directions (2023) · cites this paper as [113]
In the citing paper
“| Wu, N. et al. [113] | Semantic features | Multisemantic path neural network | ACC = 94.21 AUC = 99.52 | F++(HQ) TIMIT(LQ) |”
From Deepfake Attacks: Generation, Detection, Datasets, Challenges, and Research Directions (2023) · cites this paper as [113]
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 major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Multisemantic Path Neural Network for Deepfake Detection > 1. Introduction
“which preserves the Y channel information on the YCrCb space and compresses the CbCr information as much as possible.”
§ Multisemantic Path Neural Network for Deepfake Detection > 3.2. Compressed Videos Analysis.
“it is desirable to retain as much chromaticity information as possible and compress brightness information to save storage space.”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Multisemantic Path Neural Network for Deepfake Detection > 3.3. Multisemantic Path.
“we replace a 7 × 7 convolutional kernel with three 3 × 3 convolutional kernels”
§ Multisemantic Path Neural Network for Deepfake Detection > 4.3. Preprocessing Analysis.
“Two 5 × 5 convolution kernels replace 7 × 7 receptive fields in the macroscopic semantic path.”
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
§ Multisemantic Path Neural Network for Deepfake Detection > 4.3. Preprocessing Analysis.
“subtracted from B 1 to obtain detail image D 1 .e detail image D 2 was obtained by combining detail image D 1 and fuzzy image B 2 , and the detail image D 3 was obtained by combining detail image D 2 and fuzzy image B 3 .”
§ Multisemantic Path Neural Network for Deepfake Detection > 3.1. Multiscale Detail Enhancement.
“D 1 � I in -B 1 , D 2 � B 1 -B 2 , D 3 � B 2 -B 3 .”
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
§ Multisemantic Path Neural Network for Deepfake Detection > 3.1. Multiscale Detail Enhancement.
“where g 3 � [0.00810.01370.02200.03300.04650.06160.07660.09000.10150.09000.07660.06160.04650.04650.03300.02200.01370.0081].”
§ Multisemantic Path Neural Network for Deepfake Detection > 3.1. Multiscale Detail Enhancement.
“G 1 , G 2 , and G 3 are Gaussian kernels with respective kernel sizes of 5 × 5, 9 × 9, and 19 × 19”
Citations
Cites
Rethinking the Inception Architecture for Computer Vision
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.308 · Cited by 32,138
Unpaired image-to-image translation using cycle-consistent adversarial networks
Zhu, J.-Y., Park, T., Isola, P., & Efros, A. A. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.244 · Cited by 23,602
Xception: Deep Learning with Depthwise Separable Convolutions
Chollet, F. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.195 · Cited by 20,380 · cited 3×
“For example, the accuracy of Xception [6] was 99.”
InfluentialSqueeze-and-Excitation Networks
Hu, J., Shen, L., Albanie, S., Sun, G., & Wu, E. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2913372 · Cited by 12,157
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 2×
“Image generation techniques have developed rapidly over the past two decades, and methods such as StyleGAN [11] can produce fake images or videos that are credible to the eye.”
Progressive Growing of GANs for Improved Quality, Stability, and Variation
Karras, T., Aila, T., Laine, S., & Lehtinen, J. (2017) · arXiv · doi:10.48550/arxiv.1710.10196 · Cited by 4,727 · cited 1×
“Many models can be used, such as WGAN [13], StyleGAN, and PGGAN [14].”
MethodStarGAN: Unified Generative Adversarial Networks for Multi-domain Image-to-Image Translation
Choi, Y., Choi, M., Kim, M., Ha, J.-W., Kim, S., & Choo, J. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00916 · Cited by 4,163
A Discriminative Feature Learning Approach for Deep Face Recognition
Wen, Y., Zhang, K., Li, Z., & Qiao, Y. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46478-7_31 · Cited by 4,077 · cited 1×
“According to our investigation, the center loss function, while used in many face recognition tasks [45], does not improve performance in tasks such as handwritten number recognition.”
MethodFaceForensics++: 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 5×
“e baseline [7] has a high accuracy in uncompressed datasets, and we only evaluated our model on low- and high-compressed data.”
BackgroundMethodInfluentialDlib-ml: A Machine Learning Toolkit
King, D. E. (2009) · doi:10.5555/1577069.1755843 · Cited by 2,777 · cited 1×
“All videos were framed, and dlib [46] was used to extract the feature points of each frame of the face to help locate and clip the face area, which was expanded by 1.”
Method
Research works printed but not linked: 2
- Y. Jo, J. Park, and Sc-fegan, "Face editing generative adver- sarial network with user's sketch and color," in Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 1745-1753, Seoul, Korea, November 2019.
- L. Li, C. Chen, L. Pan, J. Zhang, and Y. Xiang, "Sok: an overview of ppg's application in authentication," 2022, [Online]. Available: https://arxiv.org/abs/2201.11291. · arXiv
Cited by
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
DeepGuardNet: A Novel CNN Architecture for DeepFake Image Detection
Amritha Devi, N., & Simon, P. (2025) · Procedia Computer Science · doi:10.1016/j.procs.2025.04.313 · Cited by 16
Shao, Y., Wang, Q., Zhang, J., Tian, H., & Zhang, Y. (2025) · International Journal of Intelligent Systems · doi:10.1155/int/3744881 · Cited by 9
“In these stages, the green framework reduces processing time to 25 and 30 units, respectively, underscoring its efciency in leveraging edge devices for localized processing [30, 36].”
Self-citationA survey on deepfake video detection datasets
Sohan, M. F., Solaiman, M., & Hasan, M. A. (2023) · Indonesian Journal of Electrical Engineering and Computer Science · doi:10.11591/ijeecs.v32.i2.pp1168-1176 · Cited by 8
Robust deepfake detection using multi-scale feature fusion
Yogarajan, G., Soundhariya, S., & Harini, R. (2025) · Multimedia Tools and Applications · doi:10.1007/s11042-025-20768-4 · Cited by 4
Compression-Aware Hybrid Framework for Deep Fake Detection in Low-Quality Video
Lagsoun, A. M., Mustapha, O., & Hedabou, M. (2025) · IEEE Access · doi:10.1109/access.2025.3592358 · Cited by 8
“Other studies [7], [35], [36], [37], and [38] used the semantic features of deep fake videos, and [7] proposed a new framework, which is the multisemantic pathway network (MSPNN) for fake face detection.”
Lagsoun, A. M., Oujaoura, M., & Jraifi, A. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-031-92602-0_39 · Cited by 4
Learning dual aggregate features for face forgery detection
Kou, Y., Jiang, Q., Zhang, J., Jin, X., Ping, W., Miao, S., & Chu, X. (2024) · Neural Computing and Applications · doi:10.1007/s00521-024-10700-6 · Cited by 3
Self-citationOptimized mixed attention‐based bidirectional gradient model for intra‐frame video forgery detection
Dhiman, N., Singh, H., & Abhishek (2026) · Journal of Forensic Sciences · doi:10.1111/1556-4029.70453
Metadata
- Authors
- Wu, N., Jin, X., Jiang, Q., Wang, P., Zhang, Y., Yao, S., & Zhou, W.
- Affiliations
- Yunnan University (CN)
- Corresponding author
- Jiang, Qian
- Venue
- Security and Communication Networks
- Publisher
- Wiley
- Published
- 2022-10-11
- Volume
- 2022
- Pages
- 1-14
- ISSN
- 1939-0114 (ISSN-L)
- 1939-0122
- OpenAlex
- W4304587257
- Semantic Scholar
- ac9d707fce3f6de5a50f64e8d8970b5e4503296c
- Cited by
- 9
- References
- 55
- Citations by year
- 2026: 1 · 2025: 5 · 2024: 1 · 2023: 2
- FWCI
- 0.65
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Locations
- Security and Communication Networks (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Funders
- National Natural Science Foundation of China (202202AD080002, 202201AT070112, 62166047, 62002313, 62002313, 62101481, 62101481, 2020SE408, 2021–2022)
- 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
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Image and Video Quality Assessment
- Keywords
- deepfake detection, face forgery detection, semantic features, video compression artifacts, center loss, face swapping
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 2026-10-07, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-07 (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.