The DOI matches its Crossref record; 45 of the 54 references it cites were found in a registry or an index.
SafeEar: Content Privacy-Preserving Audio Deepfake Detection
Open accessAbstract
Text-to-Speech (TTS) and Voice Conversion (VC) models have exhibited remarkable performance in generating realistic and natural audio. However, their dark side, audio deepfake poses a significant threat to both society and individuals. Existing countermeasures largely focus on determining the genuineness of speech based on complete original audio recordings, which however often contain private content. This oversight may refrain deepfake detection from many applications, particularly in scenarios involving sensitive information like business secrets. In this paper, we propose SafeEar, a novel framework that aims to detect deepfake audios without relying on accessing the speech content within. Our key idea is to devise a neural audio codec into a novel decoupling model that well separates the semantic and acoustic information from audio samples, and only use the acoustic information (e.g., prosody and timbre) for deepfake detection. In this way, no semantic content will be exposed to the detector. To overcome the challenge of identifying diverse deepfake audio without semantic clues, we enhance our deepfake detector with real-world codec augmentation. Extensive experiments conducted on four benchmark datasets demonstrate SafeEar's effectiveness in detecting various deepfake techniques with an equal error rate (EER) down to 2.02%. Simultaneously, it shields five-language speech content from being deciphered by both machine and human auditory analysis, demonstrated by word error rates (WERs) all above 93.93% and our user study. Furthermore, our benchmark constructed for anti-deepfake and anti-content recovery evaluation helps provide a basis for future research in the realms of audio privacy preservation and deepfake detection.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
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?
In the citing paper
“It also exhibits reduced generalisation in cross-language scenarios and decreased robustness in unknown codec conditions.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [90]
In the citing paper
“Li et al. [90] proposed the first content-privacy-preserving audio deepfake detection framework called SafeEar.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [90]
In the citing paper
“Its Transformer-based detection model incurs high computational costs, making it less suitable for resource-constrained environments.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [90]
In the citing paper
“Moreover, in adversarial scenarios where attackers have access to automatic speech recognition (ASR) models, knowledge of SafeEar’s algorithm, or even its shuffling order, SafeEar forces ASR models to exceed 90% word error rate, demonstrating its robustness in preventing unauthorised content recovery.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [90]
In the citing paper
“SafeEar achieves comparable accuracy to mainstream methods on the ASVspoof 2019 LA dataset and outperforms most baselines on the ASVspoof 2021 LA dataset.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [90]
Citations
Cites
Proceedings of the 25th international conference on Machine learning - ICML '08
(2008) · doi:10.1145/1390156 · Cited by 6,680
5th International Conference on Learning Representations (ICLR 17)
Havrylov, S., & Titov, I. (2017) · Cited by 3,503
HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units
Hsu, W.-N., Bolte, B., Tsai, Y.-H. H., Lakhotia, K., Salakhutdinov, R., & Mohamed, A. (2021) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2021.3122291 · Cited by 2,983
An Algorithm for Intelligibility Prediction of Time–Frequency Weighted Noisy Speech
Taal, C. H., Hendriks, R. C., Heusdens, R., & Jensen, J. (2011) · IEEE Transactions on Audio, Speech, and Language Processing · doi:10.1109/tasl.2011.2114881 · Cited by 2,254 · cited 1×
“(2) Short-Time Objective Intelligibility (STOI) [75]: it indicates speech signal intelligibility with its range quantified from 0 to 1 to represent the percentage of words that are correctly understood.”
Signal estimation from modified short-time Fourier transform
Griffin, D., & Lim, J. (1984) · IEEE Transactions on Acoustics, Speech, and Signal Processing · doi:10.1109/tassp.1984.1164317 · Cited by 2,095
Hybrid speech recognition with Deep Bidirectional LSTM
Graves, A., Jaitly, N., & Mohamed, A.-R. (2013) · 2013 IEEE Workshop on Automatic Speech Recognition and Understanding · doi:10.1109/asru.2013.6707742 · Cited by 1,736 · cited 2×
“There has also been a marked enhancement in the probabilistic models used in ASR systems, evolving from DNNs [32], to long-short term memory networks (LSTM) [27], and on to Con-formers [29].”
WORLD: A Vocoder-Based High-Quality Speech Synthesis System for Real-Time Applications
MORISE, M., YOKOMORI, F., & OZAWA, K. (2016) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2015edp7457 · Cited by 1,092 · cited 2×
“• WORLD [58]: WORLD is a statistical parameter-based voice synthesis system that offers fine control over the spectral and prosodic features of the synthesized audio.”
Synthesizing Obama: learning lip sync from audio
Suwajanakorn, S., Seitz, S. M., & Kemelmacher-Shlizerman, I. (2017) · ACM Transactions on Graphics · doi:10.1145/3072959.3073640 · Cited by 1,057
Neural Speech Synthesis with Transformer Network
Li, N., Liu, S., Liu, Y., Zhao, S., & Liu, M. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33016706 · Cited by 754 · cited 2×
“…main components: (1) A frontend text analysis module [77] that converts character into phoneme or linguistic features; (2) An acoustic model [19, 46, 70] that generates speech features such as Mel filter banks (FBank) or Mel-frequency cepstrum coefficient (MFCC), from either linguistic features…”
ASVspoof 2019: A large-scale public database of synthesized, converted and replayed speech
Wang, X., Yamagishi, J., Todisco, M., Delgado, H., Nautsch, A., Evans, N., Sahidullah, M., Vestman, V., Kinnunen, T., Lee, K. A., et al. (2020) · Computer Speech & Language · doi:10.1016/j.csl.2020.101114 · Cited by 505 · cited 2×
“Conversational Speech 4 This benchmark comprises four datasets, including three standard datasets—ASVspoof 2019 [68], ASVspoof 2021 [72] for deepfake detection, Librispeech [51] for content protection, and CVoiceFake we established for both aspects.”
Research works printed but not linked: 34
- Shimaa Ahmed, Amrita Roy Chowdhury, Kassem Fawaz, and Parmesh Ramanathan. 2020. Preech: A System for Privacy-Preserving Speech Transcription. In 29th USENIX Security Symposium, USENIX Security. 2703--2720.
- R. Ardila, M. Branson, K. Davis, M. Henretty, M. Kohler, J. Meyer, R. Morais, L. Saunders, F. M. Tyers, and G. Weber. 2020. Common Voice: A Massively-Multilingual Speech Corpus. In Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020). 4211--4215.
- Fahimeh Bahmaninezhad, Chunlei Zhang, and John H. L. Hansen. 2018. Convolutional Neural Network Based Speaker De-Identification. In Odyssey 2018: The Speaker and Language Recognition Workshop, 26--29 June 2018, Les Sables d'Olonne. 255--260.
- Nicholas Carlini and David A. Wagner. 2018. Audio Adversarial Examples: Targeted Attacks on Speech-to-Text. In 2018 IEEE Security and Privacy Workshops, SP Workshops 2018. 1--7.
- Ouba"ida Chouchane, Baptiste Brossier, Jorge Esteban Gamboa Gamboa, Thomas Lardy, Hemlata Tak, Orhan Ermis, Madhu R. Kamble, Jose Patino, Nicholas W. D. Evans, Melek Önen, and Massimiliano Todisco. 2021. Privacy-Preserving Voice Anti-Spoofing Using Secure Multi-Party Computation. In 22nd Annual Conference of the International Speech Communication Association, Interspeech 2021. 856--860.
- Jiangyi Deng, Fei Teng, Yanjiao Chen, Xiaofu Chen, Zhaohui Wang, and Wenyuan Xu. 2023. V-Cloak: Intelligibility-, Naturalness- & Timbre-Preserving Real-Time Voice Anonymization. In 32nd USENIX Security Symposium, USENIX Security 2023, Anaheim. 5181--5198.
- Yuchen Fan, Yao Qian, Feng-Long Xie, and Frank K. Soong. 2014. TTS Synthesis With Bidirectional LS™ Based Recurrent Neural Networks. In 15th Annual Conference of the International Speech Communication Association, INTERSPEECH 2014. 1964--1968.
- Xiaolin Hu, Kai Li, Weiyi Zhang, Yi Luo, Jean-Marie Lemercier, and Timo Gerkmann. 2021. Speech Separation Using an Asynchronous Fully Recurrent Convolutional Neural Network. In Advances in Neural Information Processing Systems 34: Annual Conference on Neural Information Processing Systems 2021, NeurIPS. 22509--22522.
- Jee-weon Jung, Hee-Soo Heo, Hemlata Tak, Hye-jin Shim, Joon Son Chung, Bong-Jin Lee, Ha-Jin Yu, and Nicholas W. D. Evans. 2022. AASIST: Audio Anti-Spoofing Using Integrated Spectro-Temporal Graph Attention Networks. In IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2022. 6367--6371.
- Tadej Justin, Vitomir Struc, Simon Dobrisek, Bostjan Vesnicer, Ivo Ipsic, and France Mihelic. 2015. Speaker De-Identification Using Diphone Recognition and Speech Synthesis. In 11th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition, FG 2015. 1--7.
- Jaehyeon Kim, Jungil Kong, and Juhee Son. 2021. Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech. In Proceedings of the 38th International Conference on Machine Learning, ICML 2021, 18--24 July 2021 (Proceedings of Machine Learning Research, Vol. 139). 5530--5540.
- Jungil Kong, Jaehyeon Kim, and Jaekyoung Bae. 2020. HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis. In Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020.
- Kundan Kumar, Rithesh Kumar, Thibault de Boissiere, Lucas Gestin, Wei Zhen Teoh, Jose Sotelo, Alexandre de Brébisson, Yoshua Bengio, and Aaron C. Courville. 2019. MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis. In Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, NeurIPS 2019, December 8--14, 2019, Vancouver, BC. 14881--14892.
- Kai Li, Xiaolin Hu, and Yi Luo. 2022. On the Use of Deep Mask Estimation Module for Neural Source Separation Systems. In 23rd Annual Conference of the International Speech Communication Association, Interspeech 2022. 5328--5332.
- Kai Li and Yi Luo. 2023. On The Design and Training Strategies for Rnn-Based Online Neural Speech Separation Systems. In IEEE International Conference on Acoustics, Speech and Signal Processing ICASSP 2023. 1--5.
- Kai Li, Runxuan Yang, and Xiaolin Hu. 2023. An Efficient Encoder-Decoder Architecture With Top-Down Attention For Speech Separation. In The Eleventh International Conference on Learning Representations, ICLR 2023.
- Menglu Li, Yasaman Ahmadiadli, and Xiao-Ping Zhang. 2022. A Comparative Study on Physical and Perceptual Features for Deepfake Audio Detection. In DDAM at MM 2022: Proceedings of the 1st International Workshop on Deepfake Detection for Audio Multimedia, Lisboa. 35--41.
- Xinfeng Li, Xiaoyu Ji, Chen Yan, Chaohao Li, Yichen Li, Zhenning Zhang, and Wenyuan Xu. 2023. Learning Normality is Enough: A Software-based Mitigation against Inaudible Voice Attacks. In 32nd USENIX Security Symposium, USENIX Security 2023, Anaheim. 2455--2472.
- Xinfeng Li, Chen Yan, Xuancun Lu, Zihan Zeng, Xiaoyu Ji, and Wenyuan Xu. 2024. Inaudible Adversarial Perturbation: Manipulating the Recognition of User Speech in Real Time. In 31st Annual Network and Distributed System Security Symposium, NDSS 2024.
- Xinfeng Li, Zhicong Zheng, Chen Yan, Chaohao Li, Xiaoyu Ji, and Wenyuan Xu. 2023 d. Towards Pitch-Insensitive Speaker Verification via Soundfield. IEEE Internet of Things Journal (2023).
- Xiaohui Liu, Meng Liu, Longbiao Wang, Kong Aik Lee, Hanyi Zhang, and Jianwu Dang. 2023. Leveraging Positional-Related Local-Global Dependency for Synthetic Speech Detection. In IEEE International Conference on Acoustics, Speech and Signal Processing ICASSP 2023. 1--5.
- Andrew Cameron Morris, Viktoria Maier, and Phil D. Green. 2004. From WER and RIL to MER and WIL: Improved Evaluation Measures For Connected Speech Recognition. In 8th International Conference on Spoken Language Processing, INTERSPEECH-ICSLP 2004. 2765--2768.
- Ahmed Mustafa, Nicola Pia, and Guillaume Fuchs. 2021. StyleMelGAN: An Efficient High-Fidelity Adversarial Vocoder with Temporal Adaptive Normalization. In IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2021. 6034--6038.
- Vassil Panayotov, Guoguo Chen, Daniel Povey, and Sanjeev Khudanpur. 2015. Librispeech: An ASR Corpus Based on Public Domain Audio Books. In 2015 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2015. 5206--5210.
- Jose Patino, Natalia A. Tomashenko, Massimiliano Todisco, Andreas Nautsch, and Nicholas W. D. Evans. 2021. Speaker Anonymisation Using the McAdams Coefficient. In 22nd Annual Conference of the International Speech Communication Association, Interspeech 2021. 1099--1103.
- Mirco Ravanelli, Titouan Parcollet, Peter Plantinga, Aku Rouhe, Samuele Cornell, Loren Lugosch, Cem Subakan, Nauman Dawalatabad, Abdelwahab Heba, Jianyuan Zhong, Ju-Chieh Chou, Sung-Lin Yeh, Szu-Wei Fu, Chien-Feng Liao, Elena Rastorgueva, Franccois Grondin, William Aris, Hwidong Na, Yan Gao, Renato De Mori, and Yoshua Bengio. 2021. SpeechBrain: A General-Purpose Speech Toolkit., Vol. abs/2106.04624 (2021). showeprint[arXiv]2106.04624 · arXiv
- Yi Ren, Chenxu Hu, Xu Tan, Tao Qin, Sheng Zhao, Zhou Zhao, and Tie-Yan Liu. 2021. FastSpeech 2: Fast and High-Quality End-to-End Text to Speech. In 9th International Conference on Learning Representations, ICLR 2021.
- Yi Ren, Yangjun Ruan, Xu Tan, Tao Qin, Sheng Zhao, Zhou Zhao, and Tie-Yan Liu. 2019. FastSpeech: Fast, Robust and Controllable Text to Speech. In Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, NeurIPS 2019, December 8--14, 2019, Vancouver, BC. 3165--3174.
- Aäron van den Oord, Oriol Vinyals, and Koray Kavukcuoglu. 2017. Neural Discrete Representation Learning. In Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, December 4--9, 2017, Long Beach, CA,. 6306--6315.
- Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention is All you Need. In Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, December 4--9, 2017, Long Beach, CA,. 5998--6008.
- Junichi Yamagishi, Xin Wang, Massimiliano Todisco, Md. Sahidullah, Jose Patino, Andreas Nautsch, Xuechen Liu, Kong Aik Lee, Tomi Kinnunen, Nicholas W. D. Evans, and Héctor Delgado. 2021. ASVspoof 2021: Accelerating Progress in Spoofed and Deepfake Speech Detection., Vol. abs/2109.00537 (2021). showeprint[arXiv]2109.00537 · arXiv
- Ryuichi Yamamoto, Eunwoo Song, and Jae-Min Kim. 2020. Parallel Wavegan: A Fast Waveform Generation Model Based on Generative Adversarial Networks with Multi-Resolution Spectrogram. In 2020 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2020. 6199--6203.
- Geng Yang, Shan Yang, Kai Liu, Peng Fang, Wei Chen, and Lei Xie. 2021. Multi-Band Melgan: Faster Waveform Generation For High-Quality Text-To-Speech. In IEEE Spoken Language Technology Workshop, SLT 2021. 492--498.
- Heiga Zen, Andrew W. Senior, and Mike Schuster. 2013. Statistical Parametric Speech Synthesis Using Deep Neural Networks. In IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2013, Vancouver, BC. 7962--7966.
Cited by
Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead
Zhang, B., Cui, H., Nguyen, V., & Whitty, M. (2025) · Sensors · doi:10.3390/s25071989 · Cited by 72
“To address this, Li et al. [90] created the CVoiceFake dataset, covering five languages and various deepfake techniques.”
From Audio Deepfake Detection to AI-Generated Music Detection–A Pathway and Overview
Li, Y., Milling, M., Specia, L., & Schuller, B. W. (2026) · IEEE Computational Intelligence Magazine · doi:10.1109/mci.2026.3721724 · Cited by 36
“Datasets focusing on spoo fi ng and speaker veri fi cation include the ASVspoof series (2015, 2019, and 2021) [68], [69], [70] and CVoice [71].”
Post-training for Deepfake Speech Detection
Ge, W., Wang, X., Liu, X., & Yamagishi, J. (2025) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru65441.2025.11434709 · Cited by 36
“The human speech portion in the dataset was excluded. includes CVoiceFake [36] and WaveFake [40], which is the vocoded version of LJSpeech [38].”
AudioTrust: Benchmarking the Multifaceted Trustworthiness of Audio Large Language ModelsPreprint
Li, K., Shen, C., Liu, Y., Han, J., Zheng, K., Zou, X., Wang, L. Z., Zhang, S., Du, X., Luo, H., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.16211 · Cited by 25
Self-citationA 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 23
“While these methods [11–14] perform well in certain controlled environments, they often struggle to generalize against increasingly sophisticated Deepfake techniques.”
PITCH: AI-assisted Tagging of Deepfake Audio Calls using Challenge-Response
Mittal, G., Jakobsson, A., Marshall, K. A., Hegde, C., & Memon, N. D. (2025) · ACM Asia Conference on Computer and Communications Security (ASIA CCS) · doi:10.1145/3708821.3733861 · Cited by 19
AWaveFormer: Audio Wavelet Transformer Network for Generalized Audio Deepfake Detection
Wang, R., Chen, Z., Wang, B., Ba, Z., & Ren, K. (2025) · IEEE Transactions on Audio, Speech and Language Processing · doi:10.1109/taslpro.2025.3611229 · Cited by 10
Apollo: Band-sequence Modeling for High-Quality Audio Restoration
Li, K., & Luo, Y. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890825 · Cited by 18
“Moreover, these methods are found to be extensively used in speech communication, particularly in telephone or internet calls, by repairing low-quality or distorted codec audio at the receiving end, thereby delivering a clearer and more natural auditory experience [4]–[9].”
Self-citationALDEN: Dual-Level Disentanglement with Meta-learning for Generalizable Audio Deepfake Detection
Xu, Y., Li, B., Li, W., Mandelli, S., Negroni, V., & Li, S. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3754741 · Cited by 9
DOTA-ME-CS: daily oriented text audio-Mandarin English-Code switching dataset
Li, Y., Wei, Z., Yu, H., Xue, J., Zhou, H., & Schuller, B. W. (2026) · Journal of Ambient Intelligence and Humanized Computing · doi:10.1007/s12652-026-05119-x · Cited by 4
“Moreover, certain deepfake audio techniques employ voice conversion (VC) [29] to alter the timbre, and other speech processing may need to meet privacy protection requirements to conceal the original timbre [30].”
Uddin, K., Khan, A. A., Farooq, M. U., & Malik, K. M. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00161 · Cited by 15
“Whereas, to preserve content privacy, Li et al. proposed SafeEar [19], a framework that decouples semantic and acoustic information, using only acoustic cues for detection.”
Gill, W., Isak, N., & Dressman, M. (2026) · IEEE Conference on Secure and Trustworthy Machine Learning (SaTML) · doi:10.1109/satml68715.2026.00015 · Cited by 3
“SafeEar [37] proposes a method for detecting deepfake audio without accessing the actual speech content.”
A Wavelet Tour of Audio Deepfake Detection
Shah, A. J., Pandey, A., Gaikwad, M. A., & Patil, H. A. (2025) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc65261.2025.11249202 · Cited by 3
Deepfake detection using deep convolutional neural network and long short-term memory
Bahadure, S., & Mane, V. M. (2026) · Multimedia Tools and Applications · doi:10.1007/s11042-026-21375-7 · Cited by 3
“For privacy-preserving audio detection, Li et al. [18] presented SafeEar, which detects voice deepfakes using acoustic features (prosody and timbre) without exposing semantic content.”
On the Optimal Selection of Mel‐Frequency Cepstral Coefficients for Voice Deepfake Detection
Falcón-López, S. A., Tobarra, L., Robles-Gómez, A., & Vargas, R. (2026) · Expert Systems · doi:10.1111/exsy.70245 · Cited by 3
“Some pipelines combine extractors like Wav2Vec2 with ResNet classifiers (Rimon et al. 2025), whereas others propose new codecs to better separate semantic and acoustic cues (Li et al. 2024).”
AFSS: Artifact-Focused Self-Synthesis for Mitigating Bias in Audio Deepfake Detection
Nguyen-Le, H.-S., Nguyen-Thanh, H.-C., Le‐Khac, N., Nguyen, D.-T., & Nguyen-Le, H.-H. (2026) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn69854.2026.11728808 · Cited by 3
“We also compare with disentanglement methods: ALDEN [16], DSVAE [26], and SafeEar [27].”
Any-to-Any Speaker Attribute Perturbation for Asynchronous Voice Anonymization
Chen, L., Guo, C., Wang, R., Lee, K. A., & Ling, Z.-H. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3592553 · Cited by 3
“Given such threats, the need for deepfake speech detection [7], [8] and voice privacy protection techniques [9] is increasing to prevent the malicious use of speaker information.”
Vulnerabilities and Defenses in Audio-Visual Attacks: A Survey From Audio to Multimodal Models
Wen, J., Wu, X., Li, Y., Jia, Y., & Zhao, S. (2026) · ACM Computing Surveys · doi:10.1145/3857219 · Cited by 4
ASTDT: an Interpretable Adaptive Spectro-Temporal Diffusion Transformer for audio deepfake detection
Wani, T. M., Qadri, S. A. A., Ashraf, A., & Amerini, I. (2025) · EURASIP Journal on Information Security · doi:10.1186/s13635-025-00217-3 · Cited by 4
“Similarly, models like SLIM [51] and SafeEar [50] introduced domain-specific representations and semantic-aware disentanglement but relied on exter-nal frozen embeddings or codec-specific augmentations, reducing transparency and increasing dependence on specific data characteristics.”
InfluentialA Comprehensive Survey of Deepfake Detection Methods, Datasets, and Future Directions
Doke, P., Dhotre, P. S., & Kharat, P. (2025) · IEEE Pune Section International Conference (PuneCon) · doi:10.1109/punecon67554.2025.11378238 · Cited by 2
“Here, we review in-depth the techniques that are considered to be at the leading edge (SOTA) in the field, and our focus is on their methodological advances, performance as reported, and limitations that are intrinsic to them [3], [5], [8], [9], [19], [20], [28], [30], [32], [34], [36].”
InfluentialMultilingual Deepfake Speech Dataset for Robust and Generalizable Detection
Mawalim, C. O., Wang, Y., Adila, A., Okada, S., & Unoki, M. (2026) · IEEE Access · doi:10.1109/access.2026.3680508 · Cited by 2
“Another large-scale multilingual dataset is CVoiceFake [21], which contains over 1.25 million bonafide and deepfake utterances across five languages, with speech data sourced from the CommonVoice dataset [26].”
InfluentialHuang, X., & Ito, A. (2025) · Applied Sciences · doi:10.3390/app15010381 · Cited by 2
Instructsing: High-Fidelity Singing Voice Generation Via Instructing Yourself
Zeng, C., Wang, C., Miao, X., Zhao, J., Jiang, Z., & Chen, Y. (2024) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt61566.2024.10832285 · Cited by 2
Dual-Channel Spoofed Speech Detection Based on Graph Attention Networks
Tan, Y., Weng, X., & Zhu, J. (2025) · Symmetry · doi:10.3390/sym17050641 · Cited by 2
“PC-DARTS [26] 4.96 0.091 RAW PC-DARTS [27] 1.77 0.0517 RawNet2 [28] 4.66 0.1294 ResNet [29] 3.72 0.119 SafeEar [30] 3.10 0.149 RW-Resnet [31] 2.98 0.082 LCNN [32] 2.76 0.078 HuRawNet2 [33] 1.96 0.139 ours 1.64 0.051 Bold denotes the best performance value for each module.”
InfluentialWhiADD: Semantic-Acoustic Fusion for Robust Audio Deepfake Detection
Cui, J., Yu, B., Wang, Q., Meng, F., & Lu, J. (2025) · Proceedings of the 33rd ACM International Conference on Multimedia · doi:10.1145/3746027.3755594 · Cited by 1
“The main challenge in generating spoofing audio is replicating the characteristics of real speakers, such as timbre and prosody, while ignoring the actual speech content [19].”
Spoofing-Aware Speaker Verification Robust Against Domain and Channel Mismatches
Zeng, C., Miao, X., Wang, X., Cooper, E., & Yamagishi, J. (2024) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt61566.2024.10832246 · Cited by 5
“Drawing inspiration from the successful utilization of multi-layer feature aggregation as seen in [7] and [44], we have also incorporated this strategy to aggregate the output of each left branch and each right branch for the anti-spoofing and ASV sub-objectives, respectively.”
Mind the Prompt: Prompting Strategies in Audio Generations for Improving Sound Classification
Ronchini, F., Wu, H.-H., Lin, W.-C., & Antonacci, F. (2025) · IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW) · doi:10.1109/icasspw65056.2025.11011113 · Cited by 1
“Privacy concerns, particularly in applications where real-world audio could infringe on individual rights [10], [11], and the limited availability of public domain datasets are significant issues [12], [13].”
Mohamed, K. M., Skaiky, A. A., Ali, H. M. S., Mohammed, A., & Mahdi, Z. A. (2025) · IEEE International Conference on Computing and Machine Intelligence (ICMI) · doi:10.1109/icmi65310.2025.11141060 · Cited by 1
“Because manual hyperparameter tuning is time-consuming, normally yielding less-than-optimal results [8,9].”
Text-Aware CLIP: A Generalized Method for Deepfake and Fake News Detection
Liu, S., Wang, Y., Guo, J., Zeng, H., Lee, I., Zhang, Y., & Li, X. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3682152 · Cited by 1
“Detection methods based on convolutional neural networks can automatically extract deep features from images and no longer rely on manually designed features[29],[30].”
Self-citationCybersecurity vulnerabilities in IoT devices
Yan, C., Ji, X., Jiang, Q., Wang, K., Wang, X., Zhu, W., Xiao, S., Li, X., & Xu, W. (2026) · Nature Reviews Electrical Engineering · doi:10.1038/s44287-026-00296-5 · Cited by 1
“The data-based approaches remedy vulnerabilities by preprocessing the algorithm’s input and output, through techniques such as data compression 86 and data desensitization 87 .”
Self-citationLightweight Multi-Cue Fusion Network for Deepfake Audio Detection
Azam, M., & Chouhan, S. A. (2026) · International Conference on Secure Cyber Computing and Communications (ICSCCC) · doi:10.1109/icsccc69031.2026.11600000 · Cited by 1
“In particular, ASVspoof 2019 introduced two widely adopted settings: Logical Access (LA) , which focuses on TTS and VC attacks, and Physical Access (PA) , which targets replay-based attacks and channel-related degradation [13].”
FLADD: Federated Learning-Based Privacy Protection for Audio Deepfake Detection
Fan, C., Tang, Z., Xue, J., Cao, J., Lv, Z., Wen, Z., & Tao, J. (2026) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2026.3727931 · Cited by 1
Comparative Study: Deepfake Detection in Tamil Speech using Hybrid Convolutional Networks
Mohamed Faheem, M., & Bharathi Mohan, G. (2025) · International Conference on Wireless Communications, Signal Processing and Networking (WiSPNET) · doi:10.1109/wispnet64060.2025.11005318
“This innovative approach addresses critical ethical concerns surrounding data privacy while maintaining effective detection capabilities, making it particularly relevant in sensitive applications such as healthcare and personal communications [5].”
Gao, M., Zhang, L., Chen, Y., Qian, F., Cui, K., Xiao, F., & Han, J. (2025) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2025.3624801
“[69], [72], [73], [74].”
Bhardwaj, A. (2025) · International Conference on Intelligent Communication Networks and Computational Techniques (ICICNCT) · doi:10.1109/icicnct66124.2025.11232782
“This approach improves the detection of deepfakes through the transformer’s ability to understand both short-and long-term dependency in speech signals [6], [9].”
InfluentialSynSpeech: A Dataset and Benchmark for Fake Speech Detection
Qiu, Q., Li, Y., Lee, L., Wang, F. L., & Yang, Z. (2025) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3743093.3771030 · Cited by 1
Combating Deepfakes : A Review of Detection Techniques and Their Limitations
Haksabanovic, A. (2025) · DiVA (Digitala Vetenskapliga Arkivet) · Cited by 1
Effect of Spoof Speech on Forensic Voice Comparison Using Deep Speaker Embeddings
Alsalihi, M. H., & Sztahó, D. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-07956-5_21
Real-Time Deepfake Audio Detection Using Lightweight CNNs
Karale, A., & Talele, P. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-13419-6_5
Audio Deepfake Detection via Dual Branch Classifier with Self-Supervised Pre-Trained Model
Cao, Y., Gao, L., & Mao, Q. (2026) · Communications in computer and information science · doi:10.1007/978-981-95-5382-2_25
Detecting and Attributing Synthetic Spanish Speech: The HISPASpoof Dataset
Risques, M., Bhagtani, K., Yadav, A. K. S., & Delp, E. J. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11464323
Real-Time Voice Liveliness Detection for Securing Voice Assistants from Deepfake Audio Attacks
Heya, T. A., & Tip, S. W. (2026) · IEEE International Conference on Semantic Computing (ICSC) · doi:10.1109/icsc67292.2026.00009
“Li et al. [17] introduced SafeEar , a privacy-preserving framework that disentangles semantic and acoustic components, enabling models to rely on prosodic cues while protecting speaker identity and linguistic content.”
Deepfake speech detection based on biological factors
Frič, M. (2026) · Brno University of Technology Digital Library (Brno University of Technology)
López, S. A. F. (2026) · UNED repository
Zhang, G., Liu, S., Zhang, D., & Tang, L. (2026) · Applied Sciences · doi:10.3390/app16178413 · Cited by 1
Nadeem, T., & Wang, S. (2026) · IEEE International Conference on Consumer Electronics - Taiwan (ICCE-Taiwan) · doi:10.1109/icce-taiwan71481.2026.11652626
Pramanik, S., Mondal, S., Bhattacharjee, S., Fouzder, T., & Roy, B. (2026) · International Conference on Power, Electronics, Communications, Computing, and Intelligent Infrastructure (PECCII) · doi:10.1109/peccii70991.2026.11662097
“English RawNet2 + SSL Front-ends ASVspoof 5 [26] 92.63% Accuracy [22] Multilingual ARPNet ASV5T1, ASVspoof21 LA, CFAD, Decro, CVoiceFake [26]–[30] ASVspoof21 LA: 1.98% EER, 0.237 t-DCF; ASV5T1: 0.272 min-DCF; CFAD: 29.56% EER; Decro: 0.40% EER; CVoiceFake: 0.98% EER [23] Arabic + English GANs…”
InfluentialA REVIEW OF MULTIMODAL DEEPFAKE FORGERY DETECTION IN AUDIOVIDEO-TEXT
Ali, H., Ghous, H., Asim, M., Rashid, T., & Akhtar, H. M. U. (2026) · Journal of Applied Linguistics and TESOL (JALT)
Li, Y., Fang, Y., Xue, J., Milling, M., & Schuller, B. W. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1928592
“Examples include ConvNext-Tiny with Whisper (Pham et al., 2024), the Self-Attention model with Hybrid Features (Huang and Pun, 2024), and the privacy-preserving model SafeEar (Li X. et al., 2024), all of which report improved spoof detection performance.”
Uncertainty-driven Multi-View Feature Fusion for Spoofed Speech Detection
Luo, X., Zhu, L., He, Z., & Wu, X. (2026) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn69854.2026.11729260
Privacy-Preserving Forged Audio Detection System Based on Semantic-Acoustic Decoupling
Wang, Z., Tian, X., Shen, Y., Liu, J., & Islam, S. M. (2026) · Communications in computer and information science · doi:10.1007/978-981-92-7537-3_16
Proof of Authorized Human Interaction for Voice AI Agents
Berisha, V., Lenz, I., Vollmer, F., & Ahmad, A. (2026) · doi:10.1145/3806096.3844842
Küçük, M., Şeker, Z. S., Çelik, B. N., & Sadık, Ş. A. (2026) · Discover Informatics · doi:10.1007/s44564-026-00042-x
Ge, Y., Jiang, J., Liu, Y., Duan, H., Zhang, S., Zhao, L., Shen, C. Q., & Wang, Q. (2026) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3830454.3832636
Metadata
- Authors
- Li, X., Li, K., Zheng, Y., Yan, C., Ji, X., & Xu, W.
- Affiliations
- Zhejiang University (CN)
- Tsinghua University (CN)
- Venue
- Proceedings of the 2024 on ACM SIGSAC Conference on Computer and Communications Security
- Publisher
- ACM
- Published
- 2024-12-02
- Pages
- 3585-3599
- ISSN
- 1543-7221 (ISSN-L)
- OpenAlex
- W4405182531
- Cited by
- 39
- References
- 54
- Citations by year
- 2026: 16 · 2025: 21 · 2024: 2
- FWCI
- 8.08
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- ACM SIGSAC Conference on Computer and Communications Security (CCS) (publishedVersion, public-domain)
- arXiv (Cornell University) (submittedVersion, cc-by-nc-nd)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Signal Processing › Music and Audio Processing
- Keywords
- audio deepfake detection, content privacy, audio forensics, deepfake audio, privacy-preserving detection
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.