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{180804,
        author = {Shreya Sakare and Chaitali Nigade and Shruti Sakare and Mayuri Salunkhe and Nilofar Mulla},
        title = {Safeguarding Society: A DeepFake Video Detection Framework},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {1},
        pages = {2780-2786},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=180804},
        abstract = {With the growing prevalence of deepfake media, the need for effective detection methods has become crucial to combat misinformation and preserve the integrity of digital content [1]. This project focuses on the development and implementation of a deepfake detection model using a combination of ResNet50v2 and Long Short-Term Memory (LSTM) networks [3]. The proposed model aims to identify and classify deepfake content through an intricate analysis of both spatial and temporal features in video. To enhance detection performance, transfer learning is employed by leveraging ResNet50v2 as the base model, which is pre-trained on large-scale datasets such as ImageNet. Instead of training the model from scratch, transfer learning enables the system to utilize the rich feature representations learned by ResNet50v2, making the detection process more efficient and accurate [6]. The ResNet50v2 component captures spatial patterns, such as facial features and inconsistencies in pixel structures, which are often indicative of manipulated media [4]. Meanwhile, LSTMs, known for their ability to process sequential data, analyze temporal features to detect irregularities in frame sequences and unnatural speech patterns—common indicators of deepfake videos [7]. The model is optimized for real-time detection, allowing it to be applied in various scenarios such as live video streams, social media content verification, and multimedia forensics [9]. Additionally, the system includes a user-friendly interface for monitoring and managing the detection process, providing detailed analysis and reports for end-users and content moderators [8].},
        keywords = {DeepFake, Long Short-Term Memory (LSTM), ResNet50v2, Transfer Learning},
        month = {June},
        }