《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (5): 1571-1577.DOI: 10.11772/j.issn.1001-9081.2022040568
所属专题: 多媒体计算与计算机仿真
收稿日期:
2022-04-24
修回日期:
2022-06-17
接受日期:
2022-06-17
发布日期:
2022-07-11
出版日期:
2023-05-10
通讯作者:
杨莹
作者简介:
许喆(1993—),男,安徽滁州人,研究实习员,硕士,CCF会员,主要研究方向:自然语言处理、时序异常检测、人脸伪造检测基金资助:
Zhe XU, Zhihong WANG, Cunyu SHAN, Yaru SUN, Ying YANG()
Received:
2022-04-24
Revised:
2022-06-17
Accepted:
2022-06-17
Online:
2022-07-11
Published:
2023-05-10
Contact:
Ying YANG
About author:
XU Zhe, born in 1993, M. S., research probationer. His research interests include natural language processing, temporal anomaly detection, face forgery detection.Supported by:
摘要:
目前有监督的人脸伪造视频检测方法需要大量标注数据。为解决视频伪造方法迭代快、种类多等现实问题,将时序异常检测中的无监督思想引入人脸伪造视频检测,将伪造视频检测任务转为无监督的视频异常检测任务,提出一种基于重构误差的无监督人脸伪造视频检测模型。首先,抽取待检测视频中连续帧的人脸特征点序列;其次,基于偏移特征、局部特征、时序特征等多粒度信息对待检测视频中人脸特征点序列进行重构;然后,计算原始序列与重构序列之间的重构误差;最后,根据重构误差的波峰频率计算得分对伪造视频进行自动检测。实验结果表明,在FaceShifter、FaceSwap等人脸视频伪造方法上,与LRNet (Landmark Recurrent Network)、Xception-c23等检测方法相比,所提方法的检测性能的曲线下方面积(AUC)最多增加了27.6%,移植性能的AUC最多增加了30.4%。
中图分类号:
许喆, 王志宏, 单存宇, 孙亚茹, 杨莹. 基于重构误差的无监督人脸伪造视频检测[J]. 计算机应用, 2023, 43(5): 1571-1577.
Zhe XU, Zhihong WANG, Cunyu SHAN, Yaru SUN, Ying YANG. Unsupervised face forgery video detection based on reconstruction error[J]. Journal of Computer Applications, 2023, 43(5): 1571-1577.
模型 | Deepfake | Face2Face | FaceShifter | FaceSwap | NeuralTextures |
---|---|---|---|---|---|
LRNet(DF) | 0.964 3 | 0.678 2 | 0.653 0 | 0.757 1 | 0.615 3 |
LRNet(NT) | 0.778 2 | 0.944 9 | 0.589 8 | 0.677 4 | 0.920 9 |
CNN-GRU-VAE | 0.914 4 | 0.632 3 | 0.752 6 | 0.837 4 | 0.525 7 |
表1 不同模型在FaceForensic++数据集上的AUC得分
Tab. 1 AUC scores of different models on FaceForensic++ dataset
模型 | Deepfake | Face2Face | FaceShifter | FaceSwap | NeuralTextures |
---|---|---|---|---|---|
LRNet(DF) | 0.964 3 | 0.678 2 | 0.653 0 | 0.757 1 | 0.615 3 |
LRNet(NT) | 0.778 2 | 0.944 9 | 0.589 8 | 0.677 4 | 0.920 9 |
CNN-GRU-VAE | 0.914 4 | 0.632 3 | 0.752 6 | 0.837 4 | 0.525 7 |
模型 | FaceForensic++ | Celeb-DF |
---|---|---|
Two-stream | 0.701 | 0.538 |
Meso4 | 0.847 | 0.548 |
MesoInception4 | 0.830 | 0.536 |
FWA | 0.801 | 0.569 |
DSP-FWA | 0.930 | 0.646 |
Xception-c23 | 0.997 | 0.653 |
Capsule | 0.966 | 0.575 |
LRNet | 0.964 | 0.569 |
CNN-GRU-VAE | 0.914 | 0.606 |
表2 通过AUC分数对不同模型的移植性能评估
Tab. 2 Transplantation performance evaluation of different models by AUC scores
模型 | FaceForensic++ | Celeb-DF |
---|---|---|
Two-stream | 0.701 | 0.538 |
Meso4 | 0.847 | 0.548 |
MesoInception4 | 0.830 | 0.536 |
FWA | 0.801 | 0.569 |
DSP-FWA | 0.930 | 0.646 |
Xception-c23 | 0.997 | 0.653 |
Capsule | 0.966 | 0.575 |
LRNet | 0.964 | 0.569 |
CNN-GRU-VAE | 0.914 | 0.606 |
模型 | 显存占用/GB | 硬盘占用/GB | 训练时间/h |
---|---|---|---|
Xception | 12 | 64 | 21 |
X-Ray | >12 | >180 | >30 |
CNN+RNN | 9 | 64 | 22.5 |
TSN | >12 | >120 | >30 |
LRNet | 3 | 1.1 | 0.2 |
CNN-GRU-VAE | 1.4 | 1.1 | 0.1 |
表3 训练成本对比
Tab. 3 Comparisons of training cost
模型 | 显存占用/GB | 硬盘占用/GB | 训练时间/h |
---|---|---|---|
Xception | 12 | 64 | 21 |
X-Ray | >12 | >180 | >30 |
CNN+RNN | 9 | 64 | 22.5 |
TSN | >12 | >120 | >30 |
LRNet | 3 | 1.1 | 0.2 |
CNN-GRU-VAE | 1.4 | 1.1 | 0.1 |
模型 | Deepfake | Face2Face | FaceShifter | FaceSwap | NeuralTexture | Celeb-DF |
---|---|---|---|---|---|---|
CNN-GRU-VAE | 0.914 4 | 0.632 3 | 0.752 6 | 0.837 4 | 0.525 7 | 0.606 3 |
不使用偏移特征 | -0.059 1 | +0.002 5 | -0.060 2 | -0.061 8 | +0.040 5 | -0.045 0 |
GRU-VAE | -0.037 2 | -0.001 4 | -0.047 7 | -0.004 2 | -0.004 8 | -0.025 4 |
CNN-GRU-AE | -0.024 8 | -0.043 1 | -0.000 3 | -0.007 0 | +0.037 3 | -0.013 6 |
表4 网络结构消融实验中不同组件对AUC分数的影响
Tab. 4 Influence of different components on AUC score in ablation study of network structure
模型 | Deepfake | Face2Face | FaceShifter | FaceSwap | NeuralTexture | Celeb-DF |
---|---|---|---|---|---|---|
CNN-GRU-VAE | 0.914 4 | 0.632 3 | 0.752 6 | 0.837 4 | 0.525 7 | 0.606 3 |
不使用偏移特征 | -0.059 1 | +0.002 5 | -0.060 2 | -0.061 8 | +0.040 5 | -0.045 0 |
GRU-VAE | -0.037 2 | -0.001 4 | -0.047 7 | -0.004 2 | -0.004 8 | -0.025 4 |
CNN-GRU-AE | -0.024 8 | -0.043 1 | -0.000 3 | -0.007 0 | +0.037 3 | -0.013 6 |
1 | MATERN F, RIESS C, STAMMINGER M. Exploiting visual artifacts to expose Deepfakes and face manipulations[C]// Proceedings of the 2019 IEEE Winter Applications of Computer Vision Workshops. Piscataway: IEEE 2019:83-92. 10.1109/wacvw.2019.00020 |
2 | AFCHAR D, NOZICK V, YAMAGISHI J, et al. MesoNet: a compact facial video forgery detection network[C]// Proceedings of the 2018 IEEE International Workshop on Information Forensics and Security. Piscataway: IEEE, 2018:1-7. 10.1109/wifs.2018.8630761 |
3 | QIAN Y Y, YIN G J, SHENG L, et al. Thinking in frequency: face forgery detection by mining frequency-aware clues[C]// Proceedings of the 2020 European Conference on Computer Vision, LNCS 12357. Cham: Springer, 2020: 86-103. 10.1007/978-3-030-58610-2_6 |
4 | LI L Z, BAO J M, ZHANG T, et al. Face X-ray for more general face forgery detection[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 5000-5009. 10.1109/cvpr42600.2020.00505 |
5 | 汤桂花,孙磊,毛秀青,等. 基于深度对齐网络的生成对抗网络伪造人脸检测[J]. 计算机应用, 2021, 41(7):1922-1927. 10.11772/j.issn.1001-9081.2020081214 |
TANG G H, SUN L, MAO X Q, et al. Generative adversarial network synthesized face detection based on deep alignment network[J]. Journal of Computer Applications, 2021, 41(7):1922-1927. 10.11772/j.issn.1001-9081.2020081214 | |
6 | 翁泽佳,陈静静,姜育刚. 基于域对抗学习的可泛化虚假人脸检测方法研究[J]. 计算机研究与发展, 2021, 58(7): 1476-1489. 10.7544/issn1000-1239.2021.20200803 |
WENG Z J, CHEN J J, JIANG Y G. On the generalization of face forgery detection with domain adversarial learning[J]. Journal of Computer Research and Development, 2021, 58(7): 1476-1489. 10.7544/issn1000-1239.2021.20200803 | |
7 | 李旭嵘,于鲲. 一种基于双流网络的Deepfakes检测技术[J]. 信息安全学报, 2020, 5(2):84-91. |
LI X R, YU K. A Deepfakes detection technique based on two-stream network[J]. Journal of Cyber Security, 2020, 5(2):84-91. | |
8 | LI Y Z, CHANG M C, LYU S W. In ictu oculi: exposing AI generated fake face videos by detecting eye blinking[C]// Proceedings of the 2018 IEEE International Workshop on Information Forensics and Security. Piscataway: IEEE, 2018: 1-7. 10.1109/wifs.2018.8630787 |
9 | YANG X, LI Y Z, LYU S W. Exposing Deep Fakes using inconsistent head poses[C]// Proceedings of the 2019 IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway: IEEE, 2019: 8261-8265. 10.1109/icassp.2019.8683164 |
10 | SUN Z K, HAN Y J, HUA Z Y, et al. Improving the efficiency and robustness of Deepfakes detection through precise geometric features[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021:3608-3617. 10.1109/cvpr46437.2021.00361 |
11 | GÜERA D, DELP E J. Deepfake video detection using recurrent neural networks[C]// Proceedings of the 15th IEEE International Conference on Advanced Video and Signal Based Surveillance. Piscataway: IEEE, 2018: 1-6. 10.1109/avss.2018.8639163 |
12 | SABIR E, CHENG J X, JAISWAL A, et al. Recurrent convolutional strategies for face manipulation detection in videos[J]. Interfaces, 2019, 3(1): 80-87. |
13 | GU Z H, CHEN Y, YAO T P, et al. Spatiotemporal inconsistency learning for DeepFake video detection[C]// Proceedings of the 29th ACM International Conference on Multimedia. New York: ACM, 2021: 3473-3481. 10.1145/3474085.3475508 |
14 | PARK D, KIM H, HOSHI Y, et al. A multimodal execution monitor with anomaly classification for robot-assisted feeding[C]// Proceedings of the 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2017: 5406-5413. 10.1109/iros.2017.8206437 |
15 | RODRIGUEZ A, BOURNE D, MASON M, et al. Failure detection in assembly: force signature analysis[C]// Proceedings of the 2010 IEEE International Conference on Automation Science and Engineering. Piscataway: IEEE, 2010: 210-215. 10.1109/coase.2010.5584452 |
16 | HUNDMAN K, CONSTANTINOU V, LAPORTE C, et al. Detecting spacecraft anomalies using LSTMs and nonparametric dynamic thresholding[C]// Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2018:387-395. 10.1145/3219819.3219845 |
17 | MALHOTRA P, RAMAKRISHNAN A, ANAND G, et al. LSTM-based encoder-decoder for multi-sensor anomaly detection[EB/OL]. (2016-07-11) [2022-06-13].. |
18 | ZONG B, SONG Q, MIN M R, et al. Deep autoencoding gaussian mixture model for unsupervised anomaly detection[EB/OL]. [2022-02-13].. |
19 | PARK D, HOSHI Y, KEMP C C. A multimodal anomaly detector for robot-assisted feeding using an LSTM-based variational autoencoder[J]. IEEE Robotics and Automation Letters, 2018, 3(3): 1544-1551. 10.1109/lra.2018.2801475 |
20 | DU M, LI F F, ZHENG G N, et al. DeepLog: anomaly detection and diagnosis from system logs through deep learning[C]// Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security. New York: ACM, 2017: 1285-1298. 10.1145/3133956.3134015 |
21 | SU Y, ZHAO Y J, NIU C H, et al. Robust anomaly detection for multivariate time series through stochastic recurrent neural network[C]// Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2019: 2828-2837. 10.1145/3292500.3330672 |
22 | KING D E. Dlib-ml: a machine learning toolkit[J]. Journal of Machine Learning Research, 2009, 10:1755-1758. |
23 | BAKER S, MATTHEWS I. Lucas-Kanade 20 years on: a unifying framework[J]. International Journal of Computer Vision, 2004, 56(3):221-255. 10.1023/b:visi.0000011205.11775.fd |
24 | KALMAN R E. A new approach to linear filtering and prediction problems[J]. Journal of Basic Engineering, 1960, 82(1):35-45. 10.1115/1.3662552 |
25 | KIM Y. Convolutional neural network for sentence classification[C]// Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. Stroudsburg, PA: ACL, 2014:1746-1751. 10.3115/v1/d14-1181 |
26 | RÖSSLER A, COZZOLINO D, VERDOLIVA L, et al. FaceForensics++: learning to detect manipulated facial images[C]// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 1-11. 10.1109/iccv.2019.00009 |
27 | LI Y Z, YANG X, SUN P, et al. Celeb-DF: a large-scale challenging dataset for DeepFake forensics[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 3204-3213. 10.1109/cvpr42600.2020.00327 |
28 | ZHOU P, HAN X T, MORARIU V I, et al. Two-stream neural networks for tampered face detection[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway: IEEE, 2017: 1831-1839. 10.1109/cvprw.2017.229 |
29 | YUEZUN L, SIWEI L. Exposing DeepFake videos by detecting face warping artifacts[C]// Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway: IEEE, 2019: 46-52. |
30 | NGUYEN H H, YAMAGISHI J, ECHIZEN I. Use of a capsule network to detect fake images and videos[EB/OL]. (2019-12-29) [2022-04-13].. 10.1109/icassp.2019.8682602 |
31 | CHOLLET F. Xception: deep learning with depthwise separable convolutions[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE 2017:1800-1807. 10.1109/cvpr.2017.195 |
32 | EKRAAM S, JIAXIN C, AYUSH J, et al. Recurrent convolutional strategies for face manipulation detection in videos[C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, Piscataway: IEEE 2019: 80-87. 10.1109/cvprw47913.2019 |
33 | WANG L M, XIONG Y J, WANG Z, et al. Temporal segment networks: towards good practices for deep action recognition[C]// Proceedings of the 2016 European Conference on Computer Vision, LNCS 9912. Cham: Springer, 2016: 20-36. |
[1] | 贾洁茹, 杨建超, 张硕蕊, 闫涛, 陈斌. 基于自蒸馏视觉Transformer的无监督行人重识别[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2893-2902. |
[2] | 夏吾吉, 黄鹤鸣, 更藏措毛, 范玉涛. 基于无监督学习和监督学习的抽取式文本摘要综述[J]. 《计算机应用》唯一官方网站, 2024, 44(4): 1035-1048. |
[3] | 江锐, 刘威, 陈成, 卢涛. 非对称端到端的无监督图像去雨网络[J]. 《计算机应用》唯一官方网站, 2024, 44(3): 922-930. |
[4] | 胡能兵, 蔡彪, 李旭, 曹旦华. 基于图池化对比学习的图分类方法[J]. 《计算机应用》唯一官方网站, 2024, 44(11): 3327-3334. |
[5] | 赵培, 乔焰, 胡荣耀, 袁新宇, 李敏悦, 张本初. 基于多域特征提取的多变量时间序列异常检测[J]. 《计算机应用》唯一官方网站, 2024, 44(11): 3419-3426. |
[6] | 彭庆媛, 王晓峰, 王军霞, 华盈盈, 唐傲, 何飞. 可满足性问题相变研究综述[J]. 《计算机应用》唯一官方网站, 2024, 44(11): 3503-3512. |
[7] | 黄梦林, 段磊, 张袁昊, 王培妍, 李仁昊. 基于Prompt学习的无监督关系抽取模型[J]. 《计算机应用》唯一官方网站, 2023, 43(7): 2010-2016. |
[8] | 葛孟婷, 万鸣华. 基于近邻监督局部不变鲁棒主成分分析的特征提取模型[J]. 《计算机应用》唯一官方网站, 2023, 43(4): 1013-1020. |
[9] | 李文博, 刘波, 陶玲玲, 罗棻, 张航. L1正则化的深度谱聚类算法[J]. 《计算机应用》唯一官方网站, 2023, 43(12): 3662-3667. |
[10] | 郑赛, 李天瑞, 黄维. 面向通信成本优化的联邦学习算法[J]. 《计算机应用》唯一官方网站, 2023, 43(1): 1-7. |
[11] | 代少升, 熊昆, 吴云铎, 肖佳伟. 多视角约束级联回归的视频人脸特征点跟踪[J]. 《计算机应用》唯一官方网站, 2022, 42(8): 2415-2422. |
[12] | 郭一阳, 于炯, 杜旭升, 杨少智, 曹铭. 基于自编码器与集成学习的离群点检测算法[J]. 《计算机应用》唯一官方网站, 2022, 42(7): 2078-2087. |
[13] | 刘睿珩, 叶霞, 岳增营. 面向自然语言处理任务的预训练模型综述[J]. 《计算机应用》唯一官方网站, 2021, 41(5): 1236-1246. |
[14] | 李旭娟, 皮建勇, 黄飞翔, 贾海朋. 基于自生成深度神经网络的4D航迹预测[J]. 计算机应用, 2021, 41(5): 1492-1499. |
[15] | 郭一村, 陈华辉. 在线哈希算法研究综述[J]. 《计算机应用》唯一官方网站, 2021, 41(4): 1106-1112. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||