AI-Based Framework for Detecting Deep-Fake Content in Social Media

  This work demonstrates a significant advancement over existing deepfake detection systems presented in the literature. Traditional models are often limited to a single input type and provide minimal interpretability, lacking features such as confidence scoring, key-frame analysis, automated alerts, and detailed reporting. Our proposed system effectively addresses these limitations through a multi-modal approach that independently analyses images, audio, and video, offering more reliable and comprehensive detection. The integration of confidence-based outputs, extraction of six representative video frames, email notifications, and downloadable reports enhances both accuracy and usability. With an expected accuracy of 90–95%, the system provides improved performance compared to the 70–85% accuracy typically reported in existing solutions. Overall, the developed system is robust, practical, and well suited for real-world applications requiring accurate and actionable deepfake identification.

[1]  A. Rössler, D. Cozzolino, L. Verdoliva, C. Riess, J. Thies, and M. Nießner, “FaceForensics++: Learning to Detect Manipulated Facial Images,” in Proc. IEEE/CVF Int. Conf. Computer Vision (ICCV), 2019, pp. 1–11.   [2]   B. Dolhansky, J. Bitton, B. Pflaum, J. Lu, R. Howes, M. Wang, and C. Canton Ferrer, “The DeepFake Detection Challenge (DFDC) Dataset,” arXiv preprint arXiv:2006.07397, 2020.   [3]  H. H. Nguyen, J. Yamagishi, and I. Echizen, “Capsule-Forensics: Using Capsule Networks to Detect Forged Images and Videos,” in Proc. IEEE Int. Conf. Acoustics, Speech and Signal Processing (ICASSP), 2019, pp. 2307–2311.   [4]  X. Yang, Y. Li, and S. Lyu, “Exposing Deep Fakes Using Inconsistent Head Poses,” in Proc. IEEE Int. Conf. Acoustics, Speech and Signal Processing (ICASSP), 2019, pp. 8261–8265.   [5]  I. Amerini, L. Galteri, R. Caldelli, and A. Del Bimbo, “DeepFake Video Detection Through Optical Flow Based CNN,” in Proc. IEEE/CVF Int. Conf. Computer Vision Workshops (ICCVW), 2019, pp. 1205–1207.   [6]  S. Agarwal, H. Farid, Y. Gu, M. He, K. Nagano, and H. Li, “Detecting DeepFake Videos by Analyzing Facial Geometry,” in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition Workshops (CVPRW), 2020, pp. 1–6.   [7] Dolhansky, B et al.(2020). DeepFake Detection Challenges (DFDC) Dataset. arXiv. A largescale dataset for testing and improving deepfake detection models   [8] Nguyen,H.H., Yamagishi, J.,&Echizen, I. (2019). Capsule-Forensics: Detection Forged Images and Videos Using Capsule Networks. ICASSP. Introduces capsule network for deepfake detection   [9]  Amerini, I., Galteri, L., et al. (2019). DeepFake Video Detection Through Optimal Flow CNN ICCVW. Uses motion and optimal flow features to identify video manipulations.   [10] Rossler A., Cozzolino.D., et al. (2019). FaceForensics++: Learning to Detect Maniputated Facial Images. ICCV. A benchmark dataset and CNN methods for detecting fake images and videos