Comparative analysis of GAN-based fusion deep neural models for fake face detection

Musiri Kailasanathan Nallakaruppan; Chiranji Lal Chowdhary; SivaramaKrishnan Somayaji; Himakshi Chaturvedi; Sujatha. R; Hafiz Tayyab Rauf; Mohamed Sharaf; Musiri Kailasanathan Nallakaruppan; Chiranji Lal Chowdhary; SivaramaKrishnan Somayaji; Himakshi Chaturvedi; Sujatha. R; Hafiz Tayyab Rauf; Mohamed Sharaf

doi:10.3934/mbe.2024071

Mathematical Biosciences and Engineering

2024, Volume 21, Issue 1: 1625-1649. doi: 10.3934/mbe.2024071

Previous Article Next Article

Research article Special Issues

Comparative analysis of GAN-based fusion deep neural models for fake face detection

Musiri Kailasanathan Nallakaruppan ^{1
,
,},
Chiranji Lal Chowdhary ¹,
SivaramaKrishnan Somayaji ¹,
Himakshi Chaturvedi ²,
Sujatha. R ³,
Hafiz Tayyab Rauf ^{4
,
,},
Mohamed Sharaf ⁵

1.
School of Computer Science Engineering and Information Systems, Vellore Institute of Technology, Vellore, Tamilnadu 632014, India
2.
School of Computer Science and Engineering, Vellore Institute of Technology, Vellore, Tamilnadu 632014, India
3.
School of Electronics and Communication Engineering, Vellore Institute of Technology, Vellore, Tamilnadu 632014, India
4.
Centre for Smart Systems, AI and Cybersecurity, Staffordshire University, Stoke-on-Trent ST4 2DE, UK
5.
Industrial Engineering Department, College of Engineering, King Saud University, P.O. Box 800, Riyadh 11421, Saudi Arabia

Academic Editor: Danilo Pelusi

Received: 05 November 2023 Revised: 06 December 2023 Accepted: 13 December 2023 Published: 02 January 2024

Fake face identity is a serious, potentially fatal issue that affects every industry from the banking and finance industry to the military and mission-critical applications. This is where the proposed system offers artificial intelligence (AI)-based supported fake face detection. The models were trained on an extensive dataset of real and fake face images, incorporating steps like sampling, preprocessing, pooling, normalization, vectorization, batch processing and model training, testing-, and classification via output activation. The proposed work performs the comparative analysis of the three fusion models, which can be integrated with Generative Adversarial Networks (GAN) based on the performance evaluation. The Model-3, which contains the combination of DenseNet-201+ResNet-102+Xception, offers the highest accuracy of 0.9797, and the Model-2 with the combination of DenseNet-201+ResNet-50+Inception V3 offers the lowest loss value of 0.1146; both are suitable for the GAN integration. Additionally, the Model-1 performs admirably, with an accuracy of 0.9542 and a loss value of 0.1416. A second dataset was also tested where the proposed Model-3 provided maximum accuracy of 86.42% with a minimum loss of 0.4054.
- GAN,
- Inception V3,
- RESNET-50,
- Mobile Net V2,
- DenseNet-201,
- VGG-16,
- Xception
Citation: Musiri Kailasanathan Nallakaruppan, Chiranji Lal Chowdhary, SivaramaKrishnan Somayaji, Himakshi Chaturvedi, Sujatha. R, Hafiz Tayyab Rauf, Mohamed Sharaf. Comparative analysis of GAN-based fusion deep neural models for fake face detection[J]. Mathematical Biosciences and Engineering, 2024, 21(1): 1625-1649. doi: 10.3934/mbe.2024071

Related Papers:

Abstract

Fake face identity is a serious, potentially fatal issue that affects every industry from the banking and finance industry to the military and mission-critical applications. This is where the proposed system offers artificial intelligence (AI)-based supported fake face detection. The models were trained on an extensive dataset of real and fake face images, incorporating steps like sampling, preprocessing, pooling, normalization, vectorization, batch processing and model training, testing-, and classification via output activation. The proposed work performs the comparative analysis of the three fusion models, which can be integrated with Generative Adversarial Networks (GAN) based on the performance evaluation. The Model-3, which contains the combination of DenseNet-201+ResNet-102+Xception, offers the highest accuracy of 0.9797, and the Model-2 with the combination of DenseNet-201+ResNet-50+Inception V3 offers the lowest loss value of 0.1146; both are suitable for the GAN integration. Additionally, the Model-1 performs admirably, with an accuracy of 0.9542 and a loss value of 0.1416. A second dataset was also tested where the proposed Model-3 provided maximum accuracy of 86.42% with a minimum loss of 0.4054.

References

[1]	I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, et al., Generative adversarial nets, Adv. Neural Inf. Process. Syst., (2014), 2672–2680. https://doi.org/10.48550/arXiv.1406.2661 doi: 10.48550/arXiv.1406.2661
[2]	M. S. Rana, M. N. Nobi, B. Murali, A. H. Sung, Deepfake detection: A systematic literature review, IEEE Access, 2022. https://doi.org/10.1109/ACCESS.2022.3154404 doi: 10.1109/ACCESS.2022.3154404
[3]	X. Deng, B. Zhao, Z. Guan, M. Xu, A new finding and unified framework for fake image detection, IEEE Signal Process. Lett., 30 (2023), 90–94. https://doi.org/ 10.1109/LSP.2023.3243770
[4]	J. Peng, B. Zou, C. Zhu, Combining external attention gan with deep convolutional neural networks for real–fake identification of luxury handbags, Vis. Comput., 39 (2023), 971–982. https://doi.org/10.1007/s00371-021-02378-x doi: 10.1007/s00371-021-02378-x
[5]	M. Zhang, L. Zhao, B. Shi, Analysis and construction of the coal and rock cutting state identification system in coal mine intelligent mining, Sci. Rep., 13 (2023), 3489. https://doi.org/10.1038/s41598-023-30617-9 doi: 10.1038/s41598-023-30617-9
[6]	Y LeCun, Y Bengio, G Hinto, DL, Nature, 521 (2015), 436–444. https://doi.org/10.1038/nature14539
[7]	H. Ding, Y. Sun, Z. Wang, N. Huang, Z. Shen, X. Cui, et al., Rgan-el: A gan and ensemble learning-based hybrid approach for imbalanced data classification, Inf. Process. Manage., 60 (2023), 103235. https://doi.org/10.1016/j.ipm.2022.103235 doi: 10.1016/j.ipm.2022.103235
[8]	V. P. Manikandan, U. Rahamathunnisa, A neural network aided attuned scheme for gun detection in video surveillance images, Image Vision Comput., 120 (2022), 104406. https://doi.org/10.1016/j.imavis.2022.104406 doi: 10.1016/j.imavis.2022.104406
[9]	J. Kolluri, R. Das, Intelligent multimodal pedestrian detection using hybrid metaheuristic optimization with DL model, Image Vision Comput., (2023), 104628. https://doi.org/10.1016/j.imavis.2023.104628 doi: 10.1016/j.imavis.2023.104628
[10]	K. Minche, Vision Transformer-Assisted Analysis of Neural Image Compression and Generation, PhD thesis, 2022. http://dx.doi.org/10.26153/tsw/43656
[11]	N. Kumari, R. Zhang, E. Shechtman, J. Y. Zhu, Ensembling off-the-shelf models for gan training, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, (2022), 10651–10662.
[12]	M. Alawadhi, W. Yan, DL from parametrically generated virtual buildings for real-world object recognition, preprint, arXiv: 2302.05283. https://doi.org/10.48550/arXiv.2302.05283
[13]	Y. Wang, C. Peng, D. Liu, N. Wang, X. Gao, Forgerynir: Deep face forgery and detection in near-infrared scenario, IEEE Trans. Inf. Forensics Secur., 17 (2022), 500–515. https://doi.org/10.1109/TIFS.2022.3146766 doi: 10.1109/TIFS.2022.3146766
[14]	H. Zhang, B. Chen, J. Wang, G. Zhao, A local perturbation generation method for gan-generated face anti-forensics, IEEE Trans. Circuits Syst. Video Technol., 2022. https://doi.org/10.1109/TCSVT.2022.3207310 doi: 10.1109/TCSVT.2022.3207310
[15]	W. Wang, X. Wang, W. Yang, J. Liu, Unsupervised face detection in the dark, IEEE Trans. Pattern Anal. Mach. Intell., 45 (2022), 1250–1266. https://doi.org/10.1109/TPAMI.2022.3152562 doi: 10.1109/TPAMI.2022.3152562
[16]	M. Khosravy, K. Nakamura, Y. Hirose, N. Nitta, N. Babaguchi, Model inversion attack by integration of deep generative models: Privacy-sensitive face generation from a face recognition system, IEEE Trans. Inf. Forensics Secur., 17 (2022), 357–372. https://doi.org/10.1109/TIFS.2022.3140687 doi: 10.1109/TIFS.2022.3140687
[17]	A. Benlamoudi, S. E. Bekhouche, M. Korichi, K. Bensid, A. Ouahabi, A. Hadid, et al., Face presentation attack detection using deep background subtraction, Sensors, 22 (2022), 3760. https://doi.org/10.3390/s22103760 doi: 10.3390/s22103760
[18]	S. H. Silva, M. Bethany, A. M. Votto, I. H. Scarff, N. Beebe, P. Najafirad, et al., Deepfake forensics analysis: An explainable hierarchical ensemble of weakly supervised models, Forensic Sci. Int.: Synergy, 4 (2022), 100217. https://doi.org/10.1016/j.fsisyn.2022.100217 doi: 10.1016/j.fsisyn.2022.100217
[19]	S. Selitskiy, N. Christou, N. Selitskaya, Using statistical and artificial neural networks meta-learning approaches for uncertainty isolation in face recognition by the established convolutional models, in Machine Learning, Optimization, and Data Science: 7th International Conference, LOD 2021, Grasmere, UK, October 4–8, 2021, Revised Selected Papers, Part II, (2022), 338–352.
[20]	A. Gowda, N. Thillaiarasu, Investigation of comparison on modified cnn techniques to classify fake face in deepfake videos, in 2022 8th International Conference on Advanced Computing and Communication Systems (ICACCS), 1 (2022), 702–707. https://doi.org/10.1109/ICACCS54159.2022.9785092
[21]	S. Rao, N. A. Shelke, A. Goel, H. Bansal, Deepfake creation and detection using ensemble DL models, in Proceedings of the 2022 Fourteenth International Conference on Contemporary Computing, (2022), 313–319. https://doi.org/10.1145/3549206.3549263
[22]	K. R. Revi, M. M. Isaac, R. Antony, M. Wilscy, Gan-generated fake face image detection using opponent color local binary pattern and DL technique, in 2022 International Conference on Connected Systems & Intelligence (CSI), (2022), 1–7. https://doi.org/10.1109/CSI54720.2022.9924077
[23]	S. Lim, M. Shin, J. Paik, Point cloud generation using deep adversarial local features for augmented and mixed reality contents, IEEE Trans. Consum. Electron., 68 (2022), 69–76. https://doi.org/10.1109/TCE.2022.3141093 doi: 10.1109/TCE.2022.3141093
[24]	Y. Zhang, A. Wang, W. Hu, DL-based consumer behavior analysis and application research, Wireless Commun. Mobile Comput., 2022 (2022). https://doi.org/10.1155/2022/4268982 doi: 10.1155/2022/4268982
[25]	W. Zheng, M. Yue, S. Zhao, S. Liu, Attention-based spatial-temporal multi-scale network for face anti-spoofing, IEEE Trans. Biom., Behav., Identity Sci., 3 (2021), 296–307. https://doi.org/10.1109/TBIOM.2021.3066983 doi: 10.1109/TBIOM.2021.3066983
[26]	S. Zhao, W. Liu, S. Liu, J. Ge, X. Liang, A hybrid-supervision learning algorithm for real-time un-completed face recognition, Comput. Electr. Eng., 101 (2022), 108090. https://doi.org/10.1016/j.compeleceng.2022.108090 doi: 10.1016/j.compeleceng.2022.108090
[27]	S. Kiruthika, V. Masilamani, Image quality assessment based fake face detection, Multimed. Tools Appl., 82 (2023), 8691–8708. https://doi.org/10.1007/s11042-021-11493-9 doi: 10.1007/s11042-021-11493-9
[28]	A. M. Luvembe, W. Li, S. Li, F. Liu, G. Xu, Dual emotion based fake news detection: A deep attention-weight update approach, Inf. Process. Manage., 60 (2023), 103354. https://doi.org/10.1016/j.ipm.2023.103354 doi: 10.1016/j.ipm.2023.103354
[29]	S. Li, W. Li, A. M. Luvembe, W. Tong, Graph contrastive learning with feature augmentation for rumor detection, IEEE Trans. Comput. Social Syst., 2023. https://doi.org/10.1109/TCSS.2023.3269303 doi: 10.1109/TCSS.2023.3269303
[30]	F. Baratzadeh, S. M. H. Hasheminejad, Customer behavior analysis to improve detection of fraudulent transactions using DL, J. AI Data Mining, 10 (2022), 87–101. https://doi.org/10.22044/jadm.2022.10124.2151 doi: 10.22044/jadm.2022.10124.2151
[31]	P. Bamoriya, G. Siddhad, H. Kaur, P. Khanna, A. Ojha, Dsb-gan: Generation of DL based synthetic biometric data, Displays, 74 (2022), 102267. https://doi.org/10.1016/j.displa.2022.102267 doi: 10.1016/j.displa.2022.102267
[32]	W. Qian, H. Li, H. Mu, Circular lbp prior-based enhanced gan for image style transfer, Int. J. Semantic Web Inf. Syst. (IJSWIS), 18 (2022), 1–15. https://doi.org/10.4018/IJSWIS.315601 doi: 10.4018/IJSWIS.315601
[33]	Z. Zhou, Y. Li, J. Li, K. Yu, Guang Kou, M. Wang, et al., Gan-siamese network for cross-domain vehicle re-identification in intelligent transport systems, IEEE Trans. Network Sci. Eng., 2022. https://doi.org/10.1109/TNSE.2022.3199919 doi: 10.1109/TNSE.2022.3199919
[34]	S. S. Reka, P. Venugopal, V. Ravi, T. Dragicevic, Privacy-based demand response modeling for residential consumers using machine learning with a cloud–fog-based smart grid environment, Energies, 16 (2023), 1655. https://doi.org/10.3390/en16041655 doi: 10.3390/en16041655
[35]	Y. Wang, H. R. Tohidypour, M. T. Pourazad, P. Nasiopoulos, V. C. M. Leung, DL-based hdr image upscaling approach for 8k uhd applications, in 2022 IEEE International Conference on Consumer Electronics (ICCE), (2022), 1–2. https://doi.org/10.1109/ICCE53296.2022.9730460
[36]	Xhlulu, 140k real and fake faces, Feb 2020.
[37]	CIPLAB @ Yonsei University, Real and fake face detection, Jan 2019.

Reader Comments

Your name:*

Email:*
© 2024 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)