Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{174455,
        author = {Yash Dewangan and Dr. Sunil B. Mane},
        title = {Robust DeepFake Video Detection Using Hybrid Model},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {10},
        pages = {4559-4566},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=174455},
        abstract = {Recent improvements in learning poses threat and challenges due to the building of fake images and live videos, which deteriorate the relationship and the credibility. Deep learning tools can drive people to come across content that they have never seen in this manner facilitating the emergence of deepfakes. Originally created for uses in the tech, commercial and Heyday is steering deepfakes to a more surgical level, retrofitting new dimension to what creators are now pumping out. Although, the improved accuracy also bring security risks because of their ubiquity and the way they are built. In this context we need models that can reliably discriminate between true positives and false positives. This paper presents recent research on deep content analysis using deep learning to solve unsolvable problems, demonstrates the advantages and limitations of the current system, and suggests future opportunities. Discriminators based on neural networks (CNN) are often used to detect changing deep news. We use the optical flow-based feature extraction method to extract time-related features, and then use them for classification, because these methods usually focus on the features of each frame of the video and cannot learn time-related information from the video, frames, Argument. Combining CNNs and RNNs with optical flow feature design forms the basis of this approach.},
        keywords = {DeepFake Detection, Deep Learning, CNN, RNN, Optical Flow},
        month = {March},
        }