Evaluating Xception with Multitask Self-Supervised Learning for AI-Generated Image and Deepfake Video Detection
DOI:
https://doi.org/10.63158/journalisi.v8i5.1864Keywords:
AI-Genereted, Multitask Self-Supervised Learning, Xception, Deepfake Detection, ClassificationAbstract
This study evaluates an Xception-based Multitask Self-Supervised Learning (MSSL) framework for detecting real and synthetic visual content across image and video modalities. Separate models were developed for CIFAKE images and FaceForensics++ videos. Beyond the primary binary-classification objective, MSSL integrates three auxiliary representation-learning tasks: rotation prediction, patch reconstruction, and artifact-map prediction. These tasks are intended to encourage the model to learn robust semantic and manipulation-related visual features. Experiments used class-balanced subsets of 5,000 CIFAKE images and 840 FaceForensics++ videos. For video classification, five frames were selected through uniform temporal sampling, and frame-level probabilities were averaged to obtain video-level predictions; a three-frame setting was also examined. A fixed decision threshold of 0.50 was defined a priori and applied consistently across both modalities. On CIFAKE, Xception+MSSL achieved 92.60% accuracy and 0.9797 ROC-AUC, outperforming the baseline Xception model, which obtained 79.60% accuracy and 0.8937 ROC-AUC. On FaceForensics++, five-frame MSSL achieved 68.25% video-level accuracy and 0.7365 ROC-AUC, compared with 62.70% and 0.6652 for Xception. The results indicate consistent within-dataset improvements, although cross-dataset and cross-generator generalization remain unverified.
Downloads
References
[1] D. S. Chinta, S. Kamineni, R. P. Chatragadda, and S. Kamepalli, “Analyzing Image Classification on AI-Generated Art Vs Human Created Art Using Deep Learning Models,” in 2024 3rd International Conference on Electrical, Electronics, Information and Communication Technologies, ICEEICT 2024, Institute of Electrical and Electronics Engineers Inc., 2024. doi: 10.1109/ICEEICT61591.2024.10718485.
[2] M. Z. Hossain, F. Uz Zaman, and M. R. Islam, “Advancing AI-Generated Image Detection: Enhanced Accuracy through CNN and Vision Transformer Models with Explainable AI Insights,” in 2023 26th International Conference on Computer and Information Technology, ICCIT 2023, Institute of Electrical and Electronics Engineers Inc., 2023. doi: 10.1109/ICCIT60459.2023.10440990.
[3] A. Singh, “A Survey of AI Text-to-Image and AI Text-to-Video Generators,” in 2023 4th International Conference on Artificial Intelligence, Robotics and Control, AIRC 2023, Institute of Electrical and Electronics Engineers Inc., 2023, pp. 32–36. doi: 10.1109/AIRC57904.2023.10303174.
[4] D. Park, H. Na, and D. Choi, “Performance Comparison and Visualization of AI-Generated-Image Detection Methods,” IEEE Access, vol. 12, pp. 62609–62627, 2024, doi: 10.1109/ACCESS.2024.3394250.
[5] J. Hall, J. Tang, and T. Cao, “Machine Learning Classification of AI and Real Images,” in 2025 Intermountain Engineering, Technology and Computing, IETC 2025, Institute of Electrical and Electronics Engineers Inc., 2025. doi: 10.1109/IETC64455.2025.11039474.
[6] L. Ma, Z. Yan, Q. Guo, Y. Liao, H. Yu, and P. Zhou, “Detecting AI-Generated Video via Frame Consistency,” in Proceedings - IEEE International Conference on Multimedia and Expo, IEEE Computer Society, 2025. doi: 10.1109/ICME59968.2025.11210049.
[7] J. Lin, Q. Zhou, and J. Li, “Dual-path AI-generated image detection: Leveraging texture-rich and texture-poor patches with global semantic features,” Digital Signal Processing: A Review Journal, vol. 171, Mar. 2025, doi: 10.1016/j.dsp.2025.105840.
[8] H. Yu, Y. Ye, J. Rong, Q. Xuan, and C. Ma, “RealHD: A High-Quality Dataset for Robust Detection of State-of-the-Art AI-Generated Images,” Association for Computing Machinery (ACM), Oct. 2025, pp. 11394–11403. doi: 10.1145/3746027.3755207.
[9] J. Park and A. Owens, “Community Forensics: Using Thousands of Generators to Train Fake Image Detectors,” in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, IEEE Computer Society, 2025, pp. 8245–8257. doi: 10.1109/CVPR52734.2025.00772.
[10] S. Gye, J. Ko, H. Shon, M. Kwon, and J. Kim, “Reducing the Content Bias for AI-generated Image Detection,” in Proceedings - 2025 IEEE Winter Conference on Applications of Computer Vision, WACV 2025, Institute of Electrical and Electronics Engineers Inc., 2025, pp. 399–408. doi: 10.1109/WACV61041.2025.00049.
[11] J. Zou, J. Wang, K. Lu, Y. Lai, K. Luo, and Z. Yu, “E2GenF: Universal AIGC image detection based on edge enhanced generalizable features,” Pattern Recognit. Lett., vol. 200, pp. 74–80, Feb. 2025, doi: 10.1016/j.patrec.2025.12.001.
[12] A. H. Javaheri, H. Motamednia, and A. Mahmoudi-Azanveh, “Enhancing the Generalization of Synthetic Image Detection Models through the Exploration of Features in Deep Detection Models,” in Iranian Conference on Machine Vision and Image Processing, MVIP, IEEE Computer Society, 2024. doi: 10.1109/MVIP62238.2024.10491148.
[13] T. Sun, “Computer vision object detection and image recognition algorithm optimization based on self-supervised learning,” IET Conference Proceedings, vol. 2025, no. 25, pp. 636–641, Oct. 2025, doi: 10.1049/icp.2025.3123.
[14] N. G. Eapen and J. George, “Exploring Self-Supervised Learning Architectures for Image Processing: Milestones and Challenges,” in 2024 IEEE 4th International Conference on ICT in Business Industry and Government, ICTBIG 2024, Institute of Electrical and Electronics Engineers Inc., 2024. doi: 10.1109/ICTBIG64922.2024.10911752.
[15] L. Wittscher, J. Diers, and C. Pigorsch, “Improving image classification robustness using self-supervision,” in Stat, John Wiley and Sons Inc, Dec. 2022. doi: 10.1002/sta4.455.
[16] C. Tang, Y. Zhou, W. Qin, Y. Zhang, R. Wu, and W. Wang, “Transformer-Driven Self-Supervised Learning for Visual Understanding: Methods and Applications,” in 2025 4th International Conference on Artificial Intelligence, Internet of Things and Cloud Computing Technology, AIoTC 2025, Institute of Electrical and Electronics Engineers Inc., 2025, pp. 155–159. doi: 10.1109/AIoTC66747.2025.11198691.
[17] M. Zou, B. Yu, Y. Zhan, S. Lyu, and K. Ma, “Semantics-Oriented Multitask Learning for DeepFake Detection: A Joint Embedding Approach,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 35, no. 10, pp. 9950–9963, 2025, doi: 10.1109/TCSVT.2025.3572508.
[18] Y. Zheng, H. Ahn, and E. Park, “Art or Artifact? Segmenting AI-Generated Images for Deeper Detection,” in ACM WDC 2025 - Proceedings of the 4th Workshop on the security implications of Deepfakes and Cheapfakes, Association for Computing Machinery, Inc, Aug. 2025, pp. 9–14. doi: 10.1145/3709022.3736544.
[19] Y. Yao, N. Desai, and M. Palaniswami, “Masked Contrastive Representation Learning for Self-Supervised Visual Pre-Training,” in 2024 IEEE 11th International Conference on Data Science and Advanced Analytics, DSAA 2024, Institute of Electrical and Electronics Engineers Inc., 2024. doi: 10.1109/DSAA61799.2024.10722789.
[20] Y. Jiang and R. A. Frick, “Same Same, But Different?: Detecting AI-Generated Videos Using Knowledge Transfer,” in ACM WDC 2025 - Proceedings of the 4th Workshop on the security implications of Deepfakes and Cheapfakes, Association for Computing Machinery, Inc, Aug. 2025, pp. 38–43. doi: 10.1145/3709022.3736547.
[21] J. Park, L. H. Park, H. E. Ahn, and T. Kwon, “Coexistence of Deepfake Defenses: Addressing the Poisoning Challenge,” IEEE Access, vol. 12, pp. 11674–11687, 2024, doi: 10.1109/ACCESS.2024.3353785.
[22] P. Joshi and V. Nivethitha, “Deep Fake Image Detection using Xception Architecture,” in 5th International Conference on Recent Trends in Computer Science and Technology, ICRTCST 2024 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 2024, pp. 533–537. doi: 10.1109/ICRTCST61793.2024.10578398.
[23] K. Yesugade and R. Jadhav, “Comparative Evaluation of Xception and DenseNet121 for Robust Deepfake Detection,” in 2025 12th International Conference on Computing for Sustainable Global Development, 2025. doi: 10.23919/INDIACom66777.2025.11115841.
[24] L. K. Gautam, “Extreme Inception Model and Its Application in Indian Food Image Classification,” in 2024 2nd DMIHER International Conference on Artificial Intelligence in Healthcare, Education and Industry, IDICAIEI 2024, Institute of Electrical and Electronics Engineers Inc., 2024. doi: 10.1109/IDICAIEI61867.2024.10842726.
[25] J. J. Bird and A. Lotfi, “CIFAKE: Image Classification and Explainable Identification of AI-Generated Synthetic Images,” IEEE Access, vol. 12, pp. 15642–15650, 2024, doi: 10.1109/ACCESS.2024.3356122.
[26] F. Chollet, “Xception: Deep Learning with Depthwise Separable Convolutions,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017. doi: 10.1109/CVPR.2017.195.
[27] S. Gidaris, P. Singh, and N. Komodakis, “Unsupervised Representation Learning by Predicting Image Rotations,” conference paper at ICLR, Mar. 2018, doi: 10.48550/arXiv.1803.07728.
[28] K. He, X. Chen, S. Xie, Y. Li, P. Dollár, and R. Girshick, “Masked Autoencoders Are Scalable Vision Learners,” Conference on Computer Vision and Pattern Recognition, 2022, doi: 10.1109/CVPR52688.2022.01553.
[29] R. Cipolla, Y. Gal, and A. Kendall, “Multi-task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics,” in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, IEEE Computer Society, Dec. 2018, pp. 7482–7491. doi: 10.1109/CVPR.2018.00781.
[30] M. Sokolova and G. Lapalme, “A systematic analysis of performance measures for classification tasks,” Inf. Process. Manag., vol. 45, no. 4, pp. 427–437, Jul. 2009, doi: 10.1016/j.ipm.2009.03.002.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Information Systems and Informatics
This work is licensed under a Creative Commons Attribution 4.0 International License.
Author's Declaration
- The Authors certify that they have read, understood, and agreed to the Journal of Information Systems and Informatics (JournalISI) submission guidelines, policies, and submission declaration. The submission has been prepared using the provided template.
- The Authors certify that all authors have approved the publication of this manuscript and that there is no conflict of interest.
- The Authors confirm that the manuscript is their original work, has not received prior publication, is not under consideration for publication elsewhere, and has not been previously published.
- The Authors confirm that all authors listed on the title page have contributed significantly to the work, have read the manuscript, attest to the validity and legitimacy of the data and its interpretation, and agree to its submission.
- The Authors confirm that the manuscript is not copied from or plagiarized from any other published work.
- The Authors declare that the manuscript will not be submitted for publication in any other journal or magazine until a decision is made by the journal editors.
- If the manuscript is finally accepted for publication, the Authors confirm that they will either proceed with publication immediately or withdraw the manuscript in accordance with the journal’s withdrawal policies.
- The Authors agree that, upon publication of the manuscript in this journal, they transfer copyright or assign exclusive rights to the publisher, including commercial rights
