International Journal of Research and Development in Engineering Sciences (IJRDES ) | A Robust Deep Learning approach for Discriminating between Authentic and AI-Generated Videos

A Robust Deep Learning approach for Discriminating between Authentic and AI-Generated Videos

The rapid advancement in processing power has empowered deep learning algorithms to produce remarkably convincing human-synthesized videos, commonly referred to as "deep fakes." This technological progression raises concerns about potential malicious applications, such as blackmail, manipulation through revenge porn, or the exploitation of political unrest using realistic face-swapping deepfakes. In response to these challenges, we propose a novel deep learning-based technique designed to reliably differentiate between genuine videos and those generated by artificial intelligence. Our approach introduces an innovative method for automatically detecting replacement and recreation deep fakes. Leveraging the capabilities of Artificial Intelligence (AI) to combat AI-driven threats, our system employs a two-step process. Initially, frame-level features are extracted using a Res-Net Convolutional Neural Network (CNN). Subsequently, these features serve as input for training a Deep Neural Network (DNN) based on a Recurrent Neural Network (RNN) architecture. To assess the effectiveness of our method, we conducted extensive evaluations on a substantial and diverse dataset. This dataset was meticulously curated by combining various sources, including Face-Forensic++, Deepfake Detection Challenge, and Celeb-DF, aiming to simulate real-time scenarios and enhance the model's performance on real-world data. Our results demonstrate the system's capability to discern alterations in videos, effectively distinguishing between deep fakes and authentic content. Furthermore, we showcase the simplicity and reliability of our approach, illustrating its competitive performance. This research contributes to the ongoing efforts in developing robust solutions for identifying and mitigating the risks associated with the proliferation of AI-generated deep fake content in various contexts.

Research Type: Cross-Sectional Research
Paper Type: Methodology Study Paper
Vol.7 , Issue 2 , Pages: 8 -12, Mar 2025
Published on: 12 Mar, 2025
Issue Type: Regular

Cite Score

:

100
No. of authors

:

75
No. of Downloads

:

43

Cite Score

:

100
No. of authors

:

75
No. of Downloads

:

43

Cite Score

:

100
No. of authors

:

75
No. of Downloads

:

43

About Authors:

Harshitha

India

Viswam Engineering College

"""

Copyright © 2025, This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC-BY-NY-SA). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Corresponding Author: Harshitha, harshithabhaskaran1@gmail.com

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

Conflict of interest: The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.