The DOI matches its Crossref record; 117 of the 117 references it cites were found in a registry or an index.
Self-Supervised Video Forensics by Audio-Visual Anomaly Detection
Open accessAbstract
Manipulated videos often contain subtle inconsistencies between their visual and audio signals. We propose a video forensics method, based on anomaly detection, that can identify these inconsistencies, and that can be trained solely using real, unlabeled data. We train an autoregressive model to generate sequences of audio-visual features, using feature sets that capture the temporal synchronization between video frames and sound. At test time, we then flag videos that the model assigns low probability. Despite being trained entirely on real videos, our model obtains strong performance on the task of detecting manipulated speech videos. Project site: https://cfeng16.github.io/audio-visual-forensics.
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 cited source doesn’t support the claim as written.
In the citing paper
“The performance of any deepfake detection method is highly dependent on the quantitative and qualitative aspects of the datasets used for development [18, 21, 25, 40, 44, 47, 49, 51, 52, 61, 63, 69, 72, 75].”
From AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset (2024) · cites this paper as [18]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Existing surveys do not examine how detection methods withstand adversarial attacks designed to evade detection [17, 21] or postprocessing operations that degrade detection performance [22-24].”
From Deepfake detection across image, video, and audio: a comprehensive survey with empirical evaluation of generalization and robustness (2026) · cites this paper as [22]
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
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 6×
“To detect anomalies, we fit an autoregressive generative model [84, 99] to sequences of synchronization features extracted from real videos, and identify low probability examples.”
MethodInfluentialAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333
(2014) · Cited by 33,419
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
Labaca-Castro, R. (2023) · Springer eBooks · doi:10.1007/978-3-658-40442-0_9 · Cited by 21,119
Representation Learning with Contrastive Predictive Coding
Oord, A. V. D., Li, Y., & Vinyals, O. (2018) · arXiv · doi:10.48550/arxiv.1807.03748 · Cited by 6,242
SGDR: Stochastic Gradient Descent with Warm Restarts
Loshchilov, I., & Hutter, F. (2016) · arXiv · doi:10.48550/arxiv.1608.03983 · Cited by 5,309
The Kinetics Human Action Video DatasetPreprint
Kay, W., Carreira, J., Simonyan, K., Zhang, B., Hillier, C., Vijayanarasimhan, S., Viola, F., Green, T., Back, T., Natsev, P., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1705.06950 · Cited by 5,004
Research works printed but not linked: 4
- radford, 2019, Language models are unsupervised multitask learners, OpenAIRE blog, 1, 9
- zhang, 2017, S3fd: Single shot scale-invariant face detector, Proceedings of the IEEE International Conference on Computer Vision, 192
- 2022, Faceswap
- zong, 2018, Deep auto-encoding gaussian mixture model for unsupervised anomaly detection, International Conference on Learning Representations
Cited by
Sorted by citations
Binding Touch to Everything: Learning Unified Multimodal Tactile Representations
Yang, F., Feng, C., Chen, Z., Park, H., Wang, D., Dou, Y., Zeng, Z., Chen, X., Gangopadhyay, S., Owens, A., et al. (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/cvpr52733.2024.02488 · Cited by 98
“Thus, we minimize the overall loss: Naturally, minimizing the contrastive objective [31, 113, 127, 145] will “pull” a visuo-tactile pair close together and “push” it away from other pairs, achieving the alignment between touch and visual embedding.”
Self-citationPVASS-MDD: Predictive Visual-Audio Alignment Self-Supervision for Multimodal Deepfake Detection
Yu, Y., Liu, X., Ni, R., Yang, S., Zhao, Y., & Kot, A. C. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3309899 · Cited by 63
“The SS Auxiliary methods [46], [47], [58] and multimodal detection method [43] are not compared in [44], thus the results of these methods are reproduced by ourselves.”
BackgroundMethodInfluentialAV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset
Cai, Z., Ghosh, S., Adatia, A. P., Hayat, M., Dhall, A., Gedeon, T., & Stefanov, K. (2024) · Proceedings of the 32nd ACM International Conference on Multimedia · doi:10.1145/3664647.3680795 · Cited by 59
SeeABLE: Soft Discrepancies and Bounded Contrastive Learning for Exposing Deepfakes
Larue, N., Vu, N., Štruc, V., Peer, P., & Christophides, V. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01921 · Cited by 59
“Another category of models [22,27,78,80] utilize temporal features for deepfake detection.”
BackgroundLiz-López, H., Keïta, M., Ahmed, A. T., Hadid, A., Huertas‐Tato, J., & Camacho, D. (2023) · Information Fusion · doi:10.1016/j.inffus.2023.102103 · Cited by 46
Integrating Audio-Visual Features For Multimodal Deepfake Detection
Muppalla, S., Jia, S., & Lyu, S. (2023) · 2023 IEEE MIT Undergraduate Research Technology Conference (URTC) · doi:10.1109/urtc60662.2023.10534969 · Cited by 46
“multimodal detectors that extract the audio-visual inconsistencies in deepfakes [14, 21, 12, 22].”
BackgroundDistributed intelligence for IoT-based smart cities: a survey
Hashem, M., Siddiqa, A., Alaba, F. A., Bilal, M., & Alhashmi, S. M. (2024) · Neural Computing and Applications · doi:10.1007/s00521-024-10136-y · Cited by 33
“T h e fr a m e w o r k t o h a nd l e e n e r gy c on s u m p ti on a nd c o m pu t a ti on a l r e s ou r ce s A no m a l y d e t ec ti on [ 135 ] C oo r d i n a ti ng m od e l s ac r o ss d i s t r i bu t e d nod e s a nd c o m b i n i ng t h e i r r e s u lt s ca n b e c h a ll e ng i ng M od e…”
Audio–visual deepfake detection using articulatory representation learning
Wang, Y., & Huang, H. (2024) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2024.104133 · Cited by 28
Generating Visual Scenes from Touch
Yang, F., Zhang, J., & Owens, A. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.02017 · Cited by 27
“Following other work in cross-modal synthesis [48, 50], we provide conditional information to our generation models through multimodal embeddings via contrastive learning [14, 66, 62, 59].”
Self-citationMIS-AVoiDD: Modality Invariant and Specific Representation for Audio-Visual Deepfake Detection
Katamneni, V. S., & Rattani, A. (2023) · International Conference on Machine Learning and Applications · doi:10.1109/icmla58977.2023.00207 · Cited by 28
“…[16] Audio-Visual 94.4 - Lip-Sync [17] Audio-Visual 83.3 0.976 AV-POI [18] Audio-Visual 86.6 0.942 VFD [19] Audio-Visual 81.5 0.861 AVForensics [20] Audio-Visual - 0.900 MultimodalTrace [21] Audio-Visual 92.9 - MIS-AVoiDD (Ours) Audio-Visual 96.2 0.973 Table II compares MIS-AVoiDD with SOTA…”
InfluentialExposing Lip-syncing Deepfakes from Mouth Inconsistencies
Datta, S. K., Jia, S., & Lyu, S. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10687902 · Cited by 33
Learning Natural Consistency Representation for Face Forgery Video Detection
Zhang, D., Xiao, Z., Li, S., Lin, F., Li, J., & Ge, S. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73010-8_24 · Cited by 33
FRADE: Forgery-aware Audio-distilled Multimodal Learning for Deepfake Detection
Nie, F., Ni, J., Zhang, J., Zhang, B., & Zhang, W. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3681672 · Cited by 20
“Meanwhile, similar to [8], we resample the audio signal to 16 𝑘𝐻𝑧 and then transform it into the log-mel spectrogram by short-time Fourier Transform with 80 mel filter banks, a hop length of 160, and a window size of 320.”
InfluentialDeCLIP: Decoding CLIP Representations for Deepfake Localization
Smeu, S., Oneaţă, E., & Oneaţă, D. (2025) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61041.2025.00025 · Cited by 24
“Our code is available at: https://github.com/bit-ml/DeCLIP .”
A review of deep learning based multimodal forgery detection for video and audio
Tan, D., Yang, Y., Niu, C., Li, S., Yang, D., & Tan, B. (2025) · Discover Applied Sciences · doi:10.1007/s42452-025-07629-3 · Cited by 19
“…VFD [73] 86.11% KoDF AVFF [76] 95.50% POI-Forensics [71] 89.90% AVAD [77] 86.90% AV-Deepfake1M MDS [75] 56.57% AP FakeAVCeleb AVFF [76] 98.50% AVAD [77] 91.30% KoDF AVFF [76] 93.10% AVAD [77] 87.60% Lav-DF UMMAFormer [78] 92.48% EER ASVspoof 2021 (LA) SSL_Anti-spoofing [79] 0.82% T23 [80] 1.32%…”
InfluentialSoK: Systematization and Benchmarking of Deepfake Detectors in a Unified Framework
Le, B. M., Kim, J., Woo, S. S., Moore, K., Abuadbba, A., & Tariq, S. (2025) · IEEE European Symposium on Security and Privacy · doi:10.1109/eurosp63326.2025.00055 · Cited by 31
“However, a systematic evaluation of recent methods with unified criteria is lacking.”
InfluentialContextual Cross-Modal Attention for Audio-Visual Deepfake Detection and Localization
Katamneni, V. S., & Rattani, A. (2024) · 2024 IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb62174.2024.10744480 · Cited by 20
“…0.962 Unsupervised Multi-modal Deepfake Detection [57] 0.968 - Audio-visual person-of-interest [18] 0.946 0.850 Audio-visual anomaly detection [20] 0.942 0.945 Multi-modal contrastive learning [36] 0.978 0.965 PVAS-MDD [62] 0.965 0.948 Novel Smart Deepfake Detection [19] 0.954 0.960…”
BackgroundInfluentialSemantics-Oriented Multitask Learning for DeepFake Detection: A Joint Embedding Approach
Zou, M., Yu, B., Zhan, Y., Lyu, S., & Ma, K. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3572508 · Cited by 16
“…[70] and vision Transformers [32], [71], [72], along with advanced training strategies such as adversarial training [25], [73], test-time training [74], graph learning [34], [36], contrastive learning [33], [38], and one-class classification [75], have also been explored for DeepFake detection.”
Increasing Importance of Joint Analysis of Audio and Video in Computer Vision: A Survey
Shahabaz, A., & Sarkar, S. (2024) · IEEE Access · doi:10.1109/access.2024.3391817 · Cited by 13
“…[186], [187], [188], [189], [190], [191], saliency prediction [59], [66], [76], [192] and other tasks that uses information from audio [4], [75], [79], [122], [158], [193], [194], [195], [196] to enhance pure-video-based classification, such as for a moving car, dog barking, or cooking an omelet.”
Background1M-Deepfakes Detection Challenge
Cai, Z., Dhall, A., Ghosh, S., Hayat, M., Kollias, D., Stefanov, K., & Tariq, U. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3689145 · Cited by 14
“The latest work on deepfake detection methodologies further supports this trend [12, 16, 31, 34, 39, 52, 54].”
MethodAudio-Visual Deepfake Detection With Local Temporal Inconsistencies
Astrid, M., Ghorbel, E., & Aouada, D. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889087 · Cited by 14
Shahzad, S. A., Hashmi, A., Peng, Y., Tsao, Y., & Wang, H. (2025) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2025.3618409 · Cited by 25
Circumventing shortcuts in audio-visual deepfake detection datasets with unsupervised learning
Smeu, S., Boldisor, D.-A., Oneaţă, D., & Oneaţă, E. (2025) · 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/cvpr52734.2025.01753 · Cited by 20
“Among the first class, prior work proposed to verify that audio and visual streams align at a semantic level, for example, from the point of view of spoken content [2, 29], or at a representation level, based on the alignment of audio and video features [17, 44].”
InfluentialHiDF: A Human-Indistinguishable Deepfake Dataset
Kang, C., Jeong, S., Lee, J., Choi, D., Woo, S. S., & Han, J. (2025) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3711896.3737399 · Cited by 12
“Extensive research has been conducted on the inconsistencies between the audio-visual modalities in deepfake videos, which demonstrates the effectiveness of utilizing multiple modalities rather than relying only on a single audio or visual feature in deepfake video detection [13, 22, 27, 80, 84].”
Systematic analysis of video tampering and detection techniques
Diwan, A., Dixit, S., Subbiah, R., & Mahadeva, R. (2024) · Cogent Engineering · doi:10.1080/23311916.2024.2424466 · Cited by 8
“By jointly analyzing multiple modalities, researchers can detect sophisticated tampering techniques that involve manipulation across different modalities (Diwan et al., 2024; Feng et al., 2023).”
Sound Localization from Motion: Jointly Learning Sound Direction and Camera Rotation
Chen, Z., Qian, S., & Owens, A. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00726 · Cited by 9
“Other approaches study audio-visual sound localization [39, 61, 16, 62], source separation [56, 31, 58, 87], active speaker detection [3, 86, 5], navigation [12, 13, 10] and forensics [105, 34, 24].”
BackgroundSelf-citationEnhancing Deepfake Detection: Proactive Forensics Techniques Using Digital Watermarking
Lai, Z. (2024) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2024.059370 · Cited by 7
“Furthermore, passive detection methods are reactive and cannot prevent the advanced spread of forged facial images or videos, and their results lack robust evidence to validate facial forgery [16–20].”
Localization and detection of deepfake videos based on self-blending method
Xu, J., Liu, X., Lin, W., Shang, W., & Wang, Y. (2025) · Scientific Reports · doi:10.1038/s41598-025-88523-1 · Cited by 10
Multi-modal Deepfake Detection via Multi-task Audio-Visual Prompt Learning
Miao, H., Guo, Y., Liu, Z., & Wang, Y. (2025) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v39i1.32042 · Cited by 9
“…multi-modal deepfake detection methods are selected, including MDS (Chugh et al. 2020), AV-DFD (Zhou and Lim 2021), BA-TFD (Cai et al. 2022), MMtrace (Raza and Malik 2023), AVAD (Feng, Chen, and Owens 2023), MRDF-CE (Zou et al. 2024), AVFF (Oorloff et al. 2024), and AVGraph (Yin et al. 2024).”
InfluentialForensics-Bench: A Comprehensive Forgery Detection Benchmark Suite for Large Vision Language Models
Wang, J., Lv, C., Li, X., Dong, S., Li, H., Yao, K., Li, C., Shao, W., & Luo, P. (2025) · 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/cvpr52734.2025.00400 · Cited by 7
“[20, 25, 30, 44, 45, 78, 81, 82, 96, 97, 111, 112] have been proposed, expecting to increase the generalization capabilities of current forgery detectors.”
Statistics-Aware Audio-Visual Deepfake Detector
Astrid, M., Ghorbel, E., & Aouada, D. (2024) · Proceedings - International Conference on Image Processing · doi:10.1109/icip51287.2024.10648224 · Cited by 7
Explicit Correlation Learning for Generalizable Cross-Modal Deepfake Detection
Yu, C., Jia, S., Fu, X., Liu, J., Tian, J., Dai, J., Wang, X., Lyu, S., & Han, J. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10687814 · Cited by 8
“AVAD [7] detects cross-modal deep-fakes by predicting synchronization features pre-learned via an external audio-visual synchronization model.”
BackgroundMethodX3A: Efficient Multimodal Deepfake Detection with Score-Level Fusion
Park, C. S., Moon, B., Jeon, M., Jung, J.-W., & Woo, S. S. (2025) · ACM Symposium on Applied Computing · doi:10.1145/3672608.3707934 · Cited by 8
“Intermediate fusion models [19, 56] merge features from both modalities at the feature level, enabling better interaction be-tween data streams.”
Modality-Agnostic Deepfakes Detection
Cai, Y., Chen, P., Tian, J., Liu, J., Dai, J., Wang, X., Jia, S., Lyu, S., & Han, J. (2025) · ACM Workshop on Information Hiding and Multimedia Security (IH&MMSec) · doi:10.1145/3733102.3733133 · Cited by 6
“…[25] 2022 Cross-modal Self-supervised Learning Binary × × BA-TFD [7] 2022 Temporal Synchronization Binary × × AVoiD-DF [64] 2023 Multi-modal Joint-decoder Binary × × AVFakeNet [30] 2023 Swin Transformer Dual-Classifier ✓ × AVAD [19] 2023 Autoregressive Synchronization Binary × × PVASS [68] 2023”
Exploring the Advancements and Challenges of Deepfake Face-swap: A Survey
Bakar, S. A. R. B. S. A., Waseem, S., Omar, Z., & Bilalashfaqahmed (2026) · Multimedia Tools and Applications · doi:10.1007/s11042-026-21285-8 · Cited by 13
A Review on Generative Intelligence in Deep Learning based Network Intrusion Detection
Ali, M., Udoidiok, I., Li, F., & Zhang, J. (2024) · Cyber Awareness and Research Symposium (CARS) · doi:10.1109/cars61786.2024.10778836 · Cited by 7
“Beyond their strong performance in modeling sequential data, autoregressive models can also contribute to IDS through their generative capabilities [44].”
Audio–visual self-supervised representation learning: A survey
Alsuwat, M. S., Al-Shareef, S., & Alghamdi, M. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.129750 · Cited by 7
Evolving from Single-modal to Multi-modal Facial Deepfake Detection: Progress and Challenges
Liu, P., Tao, Q., & Zhou, J. T. (2024) · arXiv · doi:10.48550/arxiv.2406.06965 · Cited by 6
“Feng et al. [236] formulated detection as an anomaly detection task, training an audio-visual synchronization model on pristine data.”
Aravinda Reddy, P. N., Ramachandra, R., Rao, K. S., Mitra, P., & Rathod, V. (2024) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb62174.2024.10744460 · Cited by 5
“Current studies [16, 4, 9, 7] use the majority rule, ensemble voting, and fusion module to fuse audio and visual features for accurate audio-visual deepfake detection; however, these methods are unable to adapt to changing data characteristics, limited capacity to learn complex patterns, and…”
BackgroundMethodDo Deepfake Attribution Models Generalize?
Baxevanakis, S., Schinas, M., & Papadopoulos, S. (2025) · ACM International Workshop on Multimedia AI against Disinformation (MAD) · doi:10.1145/3733567.3735572 · Cited by 6
Bi-Stream Coteaching Network for Weakly-Supervised Deepfake Localization in Videos
Li, Z., Teng, Z., Zhang, B., & Jianping Fan, B. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3533906 · Cited by 5
“For cross-dataset detection, we follow established evaluation protocols by training on a partially manipulated FF ++ (c40) dataset, with the exception of AVF [50], which is pre-trained on Lip Reading datasets.”
Xiao, M., Ye, J., Liu, B., Gao, Z., Xu, K., & Wang, X. (2025) · International Conference on Multimedia Retrieval · doi:10.1145/3731715.3733429 · Cited by 4
“Unlike supervised-based methods, existing self-supervised approaches has focused predominantly on extracting intrinsic features of authentic videos for more general deepfake detection [13, 18, 29].”
InfluentialFauForensics: Boosting Audio-Visual Deepfake Detection With Facial Action Units
Wang, J., Wu, B., Liu, L., & Liu, Q. (2026) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2026.3674401 · Cited by 7
“To capture inter-modality temporal forgery traces, most research e ff orts were devoted to exploring audio-visual dissonance [33], [43], [44], [45], [46] and joint audio-visual feature learning [4], [47], [48], [49] for improving generalization across unseen manipulation methods.”
Review of AIGC passive detection and proactive protection
Liu, X., Liu, H., Zhao, Y., Ni, R., Li, X., & Guo, M. (2025) · Scientia Sinica Informationis · doi:10.1360/ssi-2025-0043 · Cited by 3
Analytic Synaptic Dynamic Scaling Balancer for Multimodal Deepfake Continual Detection
Xiao, M., Ye, J., Liu, B., Gao, Z., Xu, K., & Wang, X. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3755680 · Cited by 3
PIA: Deepfake Detection Using Phoneme-Temporal and Identity-Dynamic Analysis
Datta, S. K., Ranga, T., Sun, C., & Lyu, S. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00170 · Cited by 4
“The audio–visual detectors exploit the lip–speech mis-matches to detect deepfakes [17, 52, 54].”
Multimodal VAD: Visual Anomaly Detection in Intelligent Monitoring System via Audio-Vision-Language
Wang, D., Wang, Q., Hu, Q., & Wu, K. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3578702 · Cited by 4
“These modalities’ complementarity e ff ectively compensates for the limitations that may exist in single modality, enabling capturing anomalous behaviors and better interpreting scenes through cross-modal fusion [24], [26], [33], [34].”
A Review of Tools and Technologies to Combat Deepfakes
Erokhin, D., & Komendantova, N. (2026) · Information · doi:10.3390/info17040347 · Cited by 3
Modal-Guided Multi-Domain Inconsistency Learning for Face Forgery Detection
Guo, Z., Zhang, B., Fan, J., Teng, Z., & Jianping Fan, B. (2024) · Applied Sciences · doi:10.3390/app15010229 · Cited by 2
Wang, W.-C., Coninck, S. D., Leroux, S., & Simoens, P. (2025) · Frontiers in Robotics and AI · doi:10.3389/frobt.2024.1490718 · Cited by 2
Domain-Adaptive Representation Learning for Multimodal Deepfake Detection
Haq, I. U., & Malik, K. M. (2025) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2025.3637242 · Cited by 2
Multiple contexts and frequencies aggregation network for deepfake detection
Li, Z., Tang, W., Gao, S., Wang, Y., & Wang, S. (2026) · PLoS ONE · doi:10.1371/journal.pone.0337409 · Cited by 5
Anshul, A., Chng, E. S., & Rajan, D. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00251 · Cited by 2
“Methods such as AVAD [14] and AVTS [44] use audio-visual synchronization tasks for pre-training, while AVFF [36] employs CV-MAE [16] to jointly learn audio-visual features.”
InfluentialSelf-Supervised Audio-Visual Soundscape Stylization
Li, T., Wang, R., Huang, P.-Y., Owens, A., & Anumanchipalli, G. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72989-8_2 · Cited by 3
Self-citationNguyen-Le, H.-H., Tran, V.-T., Nguyen, D.-T., & Le-Khac, N.-A. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-026-11608-4 · Cited by 28
Anshul, A., Gopal, S., Rajan, D., & Chng, E. S. (2025) · 2025 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.01283 · Cited by 1
“[10], we selected the maximum offset for a temporal shift, τ to be 15, thus generating 31 scores for each frame.”
GrDT: Towards Robust Deepfake Detection Using Geometric Representation Distribution and Texture
Xie, H., He, H., Fu, B., & Sanchez, V. (2025) · IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW) · doi:10.1109/wacvw65960.2025.00083 · Cited by 1
HTMDF-DD: Hybrid triple modality based spatial–temporal features early fusion for deepfake detection
Das, A. K., & Naskar, R. (2025) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2025.110860 · Cited by 1
Self-Supervised CLIP-Based Image Recognition and Analysis for Electronic Data Forensics
Wang, Y. (2025) · IEEE Access · doi:10.1109/access.2025.3648849 · Cited by 1
“In addition, Feng et al. modeled audio visual temporal inconsistencies [41], Chen et al. incorporated forgery diversity modeling and configuration prediction [42], and Wei et al.’s MaskFeat introduced a masked feature prediction mechanism—together illustrating the diverse directions of SSL research…”
A Hybrid Deep Learning Framework for Real and Deepfake Voice Detection
Murty, M., Tomar, S., & Koolagudi, S. G. (2026) · Circuits Systems and Signal Processing · doi:10.1007/s00034-025-03464-4 · Cited by 2
“Additionally, [13] presented a self-supervised method for video forensics by identifying audio-visual irregularities and utilizing discrepancies between modalities to detect modified information without requiring labeled data.”
Nie, F., Ni, J., Zhang, J., Zhang, B., Zhang, W. J., & Li, B. (2026) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2026.3673104 · Cited by 2
“Following [14], [20], facial regions are extracted from the sampled frames by MTCNN [41].”
InfluentialVTFAformer: A visual–tactile fusion network for object attribute recognition
Yi, Z., Chen, G., Wang, H., Li, L., Li, B., & Qi, F. (2026) · Information Processing & Management · doi:10.1016/j.ipm.2026.104813 · Cited by 1
AVIR: Audio-Visual Information Refinement for Robust Multimodal Deepfake Detection
Liu, X., Yang, Y. Y., Li, X., Zhao, G., & Zhao, Y. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3691216 · Cited by 2
“In order to ensure a fair comparison, we conduct training for the open-source methods [17] on our experimental datasets, refraining from the direct utilization of publicly available pre-trained models.”
InfluentialAnomaly Detection in Voice Conversations Through Spectral Features
Akilandeswari., T., Balaji, N., Paidimarri, N., & Yadav, G. N. (2024) · International Conference on Sustainable Computing and Smart Systems (ICSCSS) · doi:10.1109/icscss60660.2024.10625647 · Cited by 1
AI vs. AI: Deep Learning Approaches for Detecting Deepfake Content
Kulkarni, S., Vishwakarma, D. K., & Ranga, V. (2024) · International Conference on Advances in Computing, Communication and Networking (ICAC2N) · doi:10.1109/icac2n63387.2024.10895362 · Cited by 1
“Overall, the paper presents a huge step forward in video forensics by offering a robust and generalizable solution for detecting video manipulations from accurate, unlabeled data [9].”
Exposing Audio-Visual Forgeries in Frequency Domain
Wan, Y., Wang, J., Cui, J., & Sun, Y. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-1068-6_23
Adaptive Contribution Modulation For Multi-Modal Manipulation Media Detection and Grounding
Li, Y., & Leng, B. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890433
“Previous image manipulation detection methods [15], [16] and text manipulation detection methods [17], [18] have focused on a single modality, resulting in difficult to cope with complex multi-modal manipulations [19]–[25].”
Attention-Guided Contrastive Masked Autoencoders for Self-supervised Cross-Modal Biometric Matching
Zhang, H., Wang, J., Shi, H., Yu, Z., & Yin, L. (2025) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-031-89360-5_10
Representation Alignment For Deepfake Detection
Li, Z., Tang, W., Gao, S., Wang, Y., & Wang, S. (2025) · International Conference on Computing and Artificial Intelligence (ICCAI) · doi:10.1109/iccai66501.2025.00109
Convolutional Neural Networks for Robust Deepfake Detection in Digital Photography
Kundu, S., Bera, A., Neogi, S. K., Jaiswal, N., Ghosh, P., Das, S., & Mondal, M. A. (2025) · International Conference on Computing Technologies (ICOCT) · doi:10.1109/icoct64433.2025.11118479
DualStreamNet : Robust Audio-Video Deep Fake Media Detection Using Complimentary Information Fusion
Sheeranali, S., Ramachandra, R., & Venkatesh, S. K. (2025) · International Conference on Information Fusion (FUSION) · doi:10.23919/fusion65864.2025.11124073
“In [13], a self-supervised approach was proposed using an autoregressive model to capture the temporal synchronization between the frames and the sound of the sample.”
Improving Generalization in Deepfake Detection with Face Foundation Models and Metric Learning
Mylonas, S., & Papadopoulos, S. (2025) · doi:10.1145/3746275.3762208
“Although some recent works have explored temporal modeling [17, 18, 58, 64] or audio-visual cues [14, 28, 43] to improve robustness, our approach focuses on a more efficient, frame-based and vision-only setup.”
Tell Fake from Real: Temporal Forgery Localization with Fake-Real Guided Cross-Modal-Reconstruction
Yang, X., Zhou, Y., Xu, Y., Zou, Z., Ge, C., & He, Z. (2025) · Lecture notes in computer science · doi:10.1007/978-981-95-4381-6_20
Boosting Video Deepfake Detection via Quality-Aware and Multi-scale Audio-Visual Alignment
Chen, J., Che, X., Li, Q., & Zhao, Q. (2025) · Communications in computer and information science · doi:10.1007/978-981-95-4109-6_18
Frequency-Enhanced Multi-Modal Consistency Learning for Audio-Visual Deepfake Detection
Zheng, X., Zhang, D., & Sun, C. (2025) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3743093.3771003
“Following the prior work [11], we partition the FakeAVCeleb dataset into five categories based on the manipulation method: RVFA (Real Video, Fake Audio), FVRA-WL (Fake Video, Real Audio), FVFA-WL (Fake Video, Fake Audio, Wav2Lip), FVFA-FS (Fake Video, Fake Audio, FaceSwap), and FVFA-GAN (Fake…”
Audio-Visual Deepfake Detection via Multi-Teacher Distillation of Content and Semantic Consistency
Sun, B., Du, H., Fu, Y., Dong, T., & Lu, W. (2025) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip67698.2025.11396833
“AVoiD-DF [14] 83.70 89.20 AVAD [16] 94.20 94.50 MRDF-CE [31] 94.05 92.43 PVASS-MDD [15] 95.70 97.30 Baseline [18] – 98.86 MMMS-BA [33] 97.90 98.90 AVFF [17] 98.60 99.10 AVCS 98.26 99.64”
InfluentialCMRRL-Net: Cross-Modal Reconstruction Residual Learning for Audio–Visual Deepfake DetectionPreprint
Pandey, M., Koundal, D., & Kumar, D. S. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6597781
Audio-Visual Deepfake Generation and Detection: An Exploratory Survey
Xu, H., An, Y., Fu, P., Fan, Z., Li, B., Liu, J., Liang, Y., & Yu, M. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463954
Are DeepFakes Realistic Enough? Exploring Semantic Mismatch as a Novel Challenge
Deshmukh, S. N., Hambarde, K., Costa, J. C., Proença, H. P. M. C., & Roxo, T. (2026) · IEEE Access · doi:10.1109/access.2026.3730281
“…of forgeries requiring simultaneous audio-video manipulation [11, 22] motivated the shift to joint audio-visual learning, where methods exploit synchronization patterns, cross-modal attention, and contrastive objectives [33, 34, 12, 7] to capture forgery traces that neither modality reveals alone.”
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
Visual Surface Wave Elastography: Revealing Subsurface Physical Properties via Visible Surface Waves
Ogren, A. C., Feng, B. T., Ahn, J., Bouman, K. L., & Daraio, C. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02454 · Cited by 1
“For example, previous work has leveraged surface vibrations [14] or sound [13, 19] in video to recover unknown information.”
LN_Net: Lightweight Non-interventionist Network for Weakly-Supervised Video Anomaly Detection
Shu, L., Zhu, T., Jiang, J., Yu, Q., Li, S. M., & Peng, Y. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-7084-3_26
Zhu, J., Fan, Y., Han, G., Shi, X., Deng, F., & Chen, J. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-026-11591-w
RA-FD: Resource-Adaptive Federated Distillation for Robust Audio-Visual Anomaly Detection
Qu, B. (2026) · International Conference on Digital Management and Information Technology (DMIT) · doi:10.1145/3808707.3808862
“For instance, Feng et al. [2] successfully utilized audio-visual synchronization features for forensics, demonstrating the critical value of audio-visual complementarity.”
Liu, W., Fu, C., Peng, H., Wu, J., Chen, X., Huang, L., Jing, X., Tian, Y., Liang, M., Wang, X., et al. (2026) · International Conference on Multimedia Retrieval · doi:10.1145/3805622.3810582
Multimodal Deepfake Detection with Quantum State Inspired Analytic Incremental Adaptability Learning
Ye, J., Xiao, M., Liu, B., Hu, H., Gao, Z., Fu, S., Xu, K., & Wang, H. (2026) · International Conference on Multimedia Retrieval · doi:10.1145/3805622.3810739
“Additionally, [10, 40] presents a self-supervised method for detecting manipulated videos through anomaly detection.”
TCFFNet: Temporal and Cross-Modal Feature Fusion Network for Deepfake Detection
Wu, Z., Yang, R., Lan, R., & Ghazali, K. H. (2026) · International Conference on Image, Signal Processing and Pattern Recognition (ISPP) · doi:10.1109/ispp69262.2026.11542863
“Rep-resentative studies model temporal coherence, identity inconsistency [18], and semantic spatiotemporal forgery patterns [19]–[21], while self-supervised pre-training has been explored to improve robustness in degraded scenarios [22], [23].”
SRG-OSyNet: A Neuro-Oscillation Inspired Audio-Visual Network for Deepfake Detection
Amin, Z., Subudhi, B. N., Nathwani, K., & Jakhetiya, V. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3705934
“Complementary developments in self-supervised learning [33] and contrastive learning and [16] improved the generalization of multimodal embed-dings by leveraging unlabeled data.”
Multimodal Consistency Anomaly Detection Method Based on MLP
Wang, K., Ma, Y., Lu, S., Li, X., & Dan, F. (2026) · Chinese Control and Decision Conference (CCDC) · doi:10.1109/ccdc69976.2026.11560279
“Therefore, effectively detecting the consistency relationship between audio and video has become an important research problem in the fields of multimodal learning and anomaly detection[3].”
Adaptive Affinity Memorization With Layer Mutation for Multimodal Deepfake Continual Detection
Xiao, M., Ye, J., Liu, B., Gao, Z., Chen, W., Li, T., Wang, H., & Xu, K. (2026) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2026.3707760
“Feng et al. [24] reformulate detection as a temporal anomaly task.”
Transfer Learning with Vision Transformer for Human-Indistinguishable Deepfake Detection
Amin, U., & Mateen, M. (2026) · International Conference on IT and Industrial Technologies (ICIT) · doi:10.1109/icit68548.2026.11577692
“AP AUC AVAD [14] 0.510 0.456 MARLIN-L [15] 0.511 0.491 FTCN [13] 0.631 0.697 EB4 (Previous SOTA) [22] 0.712 0.733 CADDM-Eb4 [12] 0.669 0.649 ViT-B/16 (Ours) 0.996 0.995”
InfluentialLLM as a forensic reasoner: Expert-guided video deepfake detection across modalities
Liu, J., Wang, J., Hou, S., Ren, M., Wu, H. B., Ma, L., & He, Z. (2026) · Pattern Recognition · doi:10.2139/ssrn.7051224
Wavelet Enhanced Lip Forgery Learning for Talking Face Generation Detection
Chai, Y., Xue, J., Ren, Y. Q., Huang, Y. N., & Liu, J. (2026) · IEEE Signal Processing Letters · doi:10.1109/lsp.2026.3716160
“This high deceptiveness challenges existing detectors [3], [4], [5], [6] designed for FS artifacts.”
Deepfake Detection Using Generative Artificial Intelligence Models
Shahapurkar, M. M., Sambrekar, K. P., & Kalghatgi, S. M. (2026) · International Conference on Intelligent Technologies (CONIT) · doi:10.1109/conit69683.2026.11620609
DATS-AV: A Dissonance-Aware Two-Stage Framework for Audio–Visual Deepfake Detection
Tonmoy, R. K., & Rattani, A. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-31936-4_3
Coro, G., Cuttano, A., & Bardelli, S. (2026) · Neural Computing and Applications · doi:10.1007/s00521-026-12319-1
Dry-Type Transformers Fault Early Warning Based on SSTN-DBSCAN
Guan, X., Zhang, G., Cheng, Y., Duan, R., & Gong, Z. (2026) · Lecture notes in electrical engineering · doi:10.1007/978-981-92-2262-9_15
Zhang, W., Fan, X., Pan, Q., & Zhou, B. (2026) · IEEE Access · doi:10.1109/access.2026.3734244
“Different modalities may exhibit temporal misalignment, modality-specific noise, or missing information in practical surveillance systems [25].”
Zafar, A., Donepudi, H., Panda, R. K., Maliye, S., Sahni, N., & Nair, R. C. (2026) · IEEE International Conference on Electronics, Computing and Communication Technologies (CONECCT) · doi:10.1109/conecct68978.2026.11700622
“We compare StreamSentry against representative unimodal and multimodal baselines, including VFD [25], AVFakeNet [26], JointAV [27], AVAD [28], and AVFF.”
Leave No Stone Unturned: Uncovering Holistic Audio-Visual Intrinsic Coherence for Deepfake Detection
Peng, J., Wang, Y., Li, Y., Kong, L., & Hong, X. (2026) · doi:10.1109/cvprf71909.2026.00639
“2 in the main paper), we further conduct cross-manipulation experiments on the FakeAVCeleb dataset, following the protocols in [23, 56, 61].”
Influential
Metadata
- Authors
- Feng, C., Chen, Z., & Owens, A.
- Affiliations
- University of Michigan (US)
- Venue
- 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2023-06
- Pages
- 10491-10503
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W4386076652
- Cited by
- 101
- References
- 117
- Citations by year
- 2026: 35 · 2025: 39 · 2024: 20 · 2023: 7
- FWCI
- 10.80
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- U.S. Department of Defense
- Defense Advanced Research Projects Agency
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Physical Sciences › Mathematics › Statistics and Probability › Benford’s Law and Fraud Detection
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Keywords
- video forensics, self-supervised learning, audio-video synchronization
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
