Detection of Morphed Facial Images Using Convolutional Neural Networks
DOI:
https://doi.org/10.63949/crinfo.v2i1.005Keywords:
- Face morphing attack, CNN, biometric security, deep learning, fake face detection, MobileNet
Abstract
Face morphing attacks are posing a threat to the contemporary biometric systems particularly in e-passports and digital identity verifications. These assaults also combine faces of several people to form synthetic images which are capable of fooling face recognition models. In order to reduce this risk, this paper develops a lightweight Convolutional Neural Network (CNN)-based model to identify morphed facial images. This approach contains preprocessing, automated feature extraction and binary classification of bona fide and morphed faces. Evaluation of the experimental outcome of sample facial data reveals that the model proposed holds high recall rates and a reasonable overall accuracy. As shown by ROC and confusion matrices, morph attacks are strongly sensitive, but there are still false positives, and the precision can still be improved by using larger datasets and more sophisticated lightweight architectures like MobileNet. On balance, the results substantiate the efficacy of the deep learning in the reliable morph attack detection, as well as its possible usage in the real-life biometric security.
Downloads
References
[1] Alkishri, W., Widyarto, S., Yousif, J. H., & Al-Bahri, M. (2023). Fake face detection based on colour textual analysis using deep convolutional neural network. Journal of Internet Services and Information Security, 13(3), 143–155.
[2] Agarwal, A., &Ratha, N. (2024). Face morphing detection in social media content. In 2024 IEEE International Conference on Image Processing (ICIP) (pp. 801–806). https://doi.org/10.1109/ICIP51287.2024.10648209
[3] Hamza, M., Tehsin, S., Humayun, M., Almufareh, M. F., &Alfayad, M. (2022). A comprehensive review of face morph generation and detection of fraudulent identities. Applied Sciences, 12(24), 12545. https://doi.org/10.3390/app122412545
[4] Hakradhar, K., Tharun, K. T., Reddy, P. S. K., Anvitha, S. S., &Thangam, S. (2025). Morphed face detection. In S. Fong, N. Dey, & A. Joshi (Eds.), ICT analysis and applications (ICT4SD 2024) (Lecture Notes in Networks and Systems, Vol. 1161). Springer. https://doi.org/10.1007/978-981-97-8602-2_46
[5] Kadiri, P., Anusha, P., Prabhu, M., Asuncion, R., Pavan, V. S., &Suman, J. V. (2024). Morphed picture recognition using machine learning algorithms. In Second International Conference on Advances in Information Technology (ICAIT) (pp. 1–6). https://doi.org/10.1109/ICAIT61638.2024.10690845
[6] Morais, P., Domingues, I., & Bernardino, J. (2025). Deep learning techniques for detecting morphed face images: A literature review. IEEE Access, 13, 105952–105981. https://doi.org/10.1109/ACCESS.2025.3578199
[7] Namis, E. M., Shaker, K., & Al-Janabi, S. (2025). Approach for detecting face morphing attacks using convolution neural network. Mesopotamian Journal of Computer Science, 2025, 83–91. https://doi.org/10.58496/MJCSC/2025/005
[8] Pokhytun, A., Mazurets, O., Molchanova, M., &Tyschenko, O. (2024). Method for neural network detecting changed images of people's faces using CNN. https://elar.khmnu.edu.ua/handle/123456789/16938
[9] Razaq, I. S. (2023). Improved face morphing attack detection method using PCA and convolutional neural network. Karbala International Journal of Modern Science, 9(2), 15. https://doi.org/10.33640/2405-609X.3298
[10] Seibold, C., Samek, W., Hilsmann, A., &Eisert, P. (2020). Accurate and robust neural networks for face morphing attack detection. Journal of Information Security and Applications, 53, 102526. https://doi.org/10.1016/j.jisa.2020.102526
[11] Venkatesh, S., Ramachandra, R., Raja, K., & Busch, C. (2021). Face morphing attack generation and detection: A comprehensive survey. IEEE Transactions on Technology and Society, 2(3), 128–145. https://doi.org/10.1109/TTS.2021.3066254
