Deephoax Image Detection based on Deep Learning using Convolutional Neural Network Architectures

SY Yuliani; Nasywa Naura  Aulia

doi:10.29207/resti.v10i7403

Authors

SY Yuliani Universitas Multimedia Nusantara
Nasywa Naura Aulia Universitas Multimedia Nusantara

DOI:

https://doi.org/10.29207/resti.v10i7403

Keywords:

convolutional neural network, deephoax detection, image classification, mobilenet, Xception

Abstract

The rapid growth of generative artificial intelligence has increased the spread of deephoax images, creating significant challenges for digital security, public trust, and the reliability of online information. This study evaluates the performance of MobileNet and Xception architectures for classifying facial images into fact and hoax categories while analyzing different feature representation levels within the Xception architecture through an ablation-based fine-tuning strategy. The proposed framework consists of data preparation, model architecture design, training procedures, and performance evaluation using confusion matrix analysis and metrics such as accuracy, precision, recall, and F1-score. Two publicly available datasets, DeepDetect2025 and FF-GenAI were utilized to evaluate the robustness of the proposed models. Experimental results show that Xception consistently outperformed MobileNet across all evaluation metrics, with the Middle Level Layer configuration achieving the best performance at 99.71% accuracy and F1-score. The findings indicate that intermediate feature representations are the most effective for capturing structural inconsistencies, texture irregularities, and synthetic blending artifacts commonly found in AI-generated facial images. In contrast, low-level representations were less discriminative, while highly abstract semantic representations slightly reduced sensitivity to localized manipulation artifacts. Overall, this study demonstrates the effectiveness of Xception-based feature refinement for deephoax image detection and contributes to AI-based approaches for digital content verification, cybersecurity, and visual misinformation mitigation.

Downloads

Download data is not yet available.

References

[1] A. Mulahuwaish et al., “A survey of social cybersecurity: Techniques for attack detection, evaluations, challenges, and future prospects,” May 01, 2025, Elsevier B.V. doi: https://doi.org/10.1016/j.chbr.2025.100668.

[2] S. Y. Yakub et al., “Cyber Security for Hoax News Detection with Similarity Algorithm,” in 2024 9th International Conference on Informatics and Computing, ICIC 2024, Institute of Electrical and Electronics Engineers Inc., 2024. doi: https://doi.org/10.1109/ICIC64337.2024.10956546.

[3] E. Broda and J. Strömbäck, “Misinformation, disinformation, and fake news: lessons from an interdisciplinary, systematic literature review,” Ann. Int. Commun. Assoc., vol. 48, no. 2, pp. 139–166, 2024, doi: https://doi.org/10.1080/23808985.2024.2323736.

[4] S. Y. Yakub et al., “Deep Learning for Cyber Security: Deephoax Detection,” in 2025 2nd International Conference on Advanced Innovations in Smart Cities, ICAISC 2025, Institute of Electrical and Electronics Engineers Inc., 2025. doi: https://doi.org/10.1109/ICAISC64594.2025.10959570.

[5] S. Alanazi and S. Asif, “Exploring deepfake technology: creation, consequences and countermeasures,” Human-Intelligent Systems Integration, vol. 6, no. 1, pp. 49–60, Dec. 2024, doi: https://doi.org/10.1007/s42454-024-00054-8.

[6] A. T. Wibowo, “Hoax and Fake News by Saracen Syndicate and the Problems for National Cyber Security,” Indonesian Journal of Counter Terrorism and National Security, vol. 1, no. 1, pp. 91–108, Jan. 2022, doi: https://doi.org/10.15294/ijctns.v1i1.56732.

[7] H. Sun, T. Zhu, Z. Zhang, D. Jin. P. Xiong, and W. Zhou, “Adversarial Attacks Against Deep Generative Models on Data: A Survey,” Dec. 2021, doi: https://doi.org/10.1109/TKDE.2021.3130903.

[8] L. Zhang et al., “X-Transfer: A Transfer Learning-Based Framework for GAN-Generated Fake Image Detection,” Jan. 2024, doi: https://doi.org/10.1109/IJCNN60899.2024.10650566.

[9] Tim Cek Fakta, “[HOAKS] Video Prabowo Menyebutkan NIK KTP Berisi Bantuan Rp 7 Juta,” Kompas, Oct. 11, 2025. Accessed: Nov. 27, 2025. [Online]. Available: https://www.kompas.com/cekfakta/read/2025/10/11/084800182/-hoaks-video-prabowo-menyebutkan-nik-ktp-berisi-bantuan-rp-7-juta

[10] M. Linawati, “Cek Fakta: Tidak Benar Video Prabowo Cairkan Bansos Rp 7 Juta per KTP,” Liputan6, Jakarta, Sep. 09, 2025. Accessed: Nov. 27, 2025. [Online]. Available: https://www.liputan6.com/cek-fakta/read/6154574/cek-fakta-tidak-benar-video-prabowo-cairkan-bansos-rp-7-juta-per-ktp#google_vignette

[11] Z. Xia, T. Qiao, M. Xu, X. Wu, L. Han, and Y. Chen, “Deepfake Video Detection Based on MesoNet with Preprocessing Module,” Symmetry (Basel)., vol. 14, no. 5, May 2022, doi: https://doi.org/10.3390/sym14050939.

[12] L. E. Demir and Y. Canbay, “Deepfake Image Detection with Transfer Learning Models,” Bitlis Eren Üniversitesi Fen Bilimleri Dergisi, vol. 14, no. 1, pp. 546–560, Mar. 2025, doi: https://doi.org/10.17798/bitlisfen.1610300.

[13] Y. Xu and S. Y. Yayilgan, “When Handcrafted Features and Deep Features Meet Mismatched Training and Test Sets for Deepfake Detection,” Sep. 2022, doi: https://doi.org/10.48550/arXiv.2209.13289.

[14] H. Huo, Y. L. Yu, and Z. H. Liu, “Facial expression recognition based on improved depthwise separable convolutional network,” Multimed. Tools Appl., vol. 82, no. 12, pp. 18635–18652, May 2023, doi: https://doi.org/10.1007/s11042-022-14066-6.

[15] Q. Jiang et al., “Robust manipulated media localization and detection based on high frequency and texture features,” Discover Computing, vol. 28, no. 1, Dec. 2025, doi: https://doi.org/10.1007/s10791-025-09500-w.

[16] B. Dhanalaxmi, B. N. Kumar, Y. Raju, and R. S. R. Channapragada, “MobileNetV3: an efficient deep learning-based feature selection and classification technique for cardiovascular disease,” Journal of Engineering and Applied Science, vol. 72, no. 1, Dec. 2025, doi: https://doi.org/10.1186/s44147-025-00654-4.

[17] “DeepDetect-2025,” Kaggle. Accessed: Dec. 14, 2025. [Online]. Available: https://www.kaggle.com/datasets/ayushmandatta1/deepdetect-2025

[18] “Flickr-Face-HQ and GenAI Dataset (FF-GenAI),” Kaggle. Accessed: Oct. 28, 2025. [Online]. Available: https://www.kaggle.com/datasets/argonautex/flickr-face-hq-and-genai-dataset-ff-genai

[19] M. Bundea and G. M. Danciu, “Pneumonia Image Classification Using DenseNet Architecture,” Information (Switzerland), vol. 15, no. 10, Oct. 2024, doi: https://doi.org/10.3390/info15100611.

[20] M. Abbasi, P. Váz, J. Silva, and P. Martins, “Comprehensive Evaluation of Deepfake Detection Models: Accuracy, Generalization, and Resilience to Adversarial Attacks,” Applied Sciences (Switzerland), vol. 15, no. 3, Feb. 2025, doi: https://doi.org/10.3390/app15031225.

[21] H. Noprisson, V. Ayumi, M. Purba, and N. Ani, “MOBILENET PERFORMANCE IMPROVEMENTS FOR DEEPFAKE IMAGE IDENTIFICATION USING ACTIVATION FUNCTION AND REGULARIZATION,” JITK (Jurnal Ilmu Pengetahuan dan Teknologi Komputer), vol. 10, no. 2, pp. 441–448, Nov. 2024, doi: https://doi.org/10.33480/jitk.v10i2.5798.

[22] A. Jagdale, V. Kubde, R. Kortikar, and N. Rajgure, “Deepfake Image Detection using Deep Learning,” International Journal of Scientific Research in Engineering and Management, 2025, doi: https://doi.org/10.55041/IJSREM46579.

[23] H. M. Lala and H. S. Ahmed, “An impact on convolutional neural networks amelioration for early detection of skin cancer,” International Journal of Intelligent Computing and Information Sciences, vol. 24, no. 2, pp. 1–17, Jun. 2024, doi: https://doi.org/10.21608/ijicis.2024.273450.1325.

[24] R. Adityatama and A. T. Putra, “Image classification of Human Face Shapes Using Convolutional Neural Network Xception Architecture with Transfer Learning,” Recursive Journal of Informatics, vol. 1, no. 2, pp. 102–109, Sep. 2023, doi: https://doi.org/10.15294/rji.v1i2.70774.

[25] R. Yao, Z. Bai, J. Tong, and K. Rezaee, “Deepfake Detection in Image Sequences: A Temporal Approach for Anomaly Detection,” International Journal of Intelligent Systems, vol. 2025, no. 1, 2025, doi: https://doi.org/10.1155/int/8566328.

[26] B. Gülmez, “A novel deep neural network model based Xception and genetic algorithm for detection of COVID-19 from X-ray images,” Ann. Oper. Res., vol. 328, no. 1, pp. 617–641, Sep. 2023, doi: https://doi.org/10.1007/s10479-022-05151-y.

[27] S. R. Dubey, S. K. Singh, and B. B. Chaudhuri, “Activation Functions in Deep Learning: A Comprehensive Survey and Benchmark,” Jun. 2022, doi: https://doi.org/10.1016/j.neucom.2022.06.111.

[28] I. Salehin and D. K. Kang, “A Review on Dropout Regularization Approaches for Deep Neural Networks within the Scholarly Domain,” Jul. 01, 2023, Multidisciplinary Digital Publishing Institute (MDPI). doi: https://doi.org/10.3390/electronics12143106.

[29] H. Kamal and M. Mashaly, “Enhanced Hybrid Deep Learning Models-Based Anomaly Detection Method for Two-Stage Binary and Multi-Class Classification of Attacks in Intrusion Detection Systems,” Algorithms, vol. 18, no. 2, Feb. 2025, doi: https://doi.org/10.3390/a18020069.

[30] O. Rainio, J. Teuho, and R. Klén, “Evaluation metrics and statistical tests for machine learning,” Sci. Rep., vol. 14, no. 1, Dec. 2024, doi: https://doi.org/10.1038/s41598-024-56706-x.

[31] C. Miller, T. Portlock, D. M. Nyaga, and J. M. O’Sullivan, “A review of model evaluation metrics for machine learning in genetics and genomics,” 2024, Frontiers Media SA. doi: https://doi.org/10.3389/fbinf.2024.1457619.

[32] C. Sweeney, E. Ennis, M. Mulvenna, R. Bond, and S. O’neill, “How Machine Learning Classification Accuracy Changes in a Happiness Dataset with Different Demographic Groups,” Computers, vol. 11, no. 5, May 2022, doi: https://doi.org/10.3390/computers11050083.