Paper Key : IRJ************813
Author: Chappidi Charitha, C.nancy
Date Published: 24 Mar 2025
Abstract
Deepfakes, which are highly realistic altered videos and images, have gained significant attention due to their potential for both positive and harmful applications. Malicious uses, such as spreading fake news, creating celebrity impersonations, and engaging in financial fraud, are increasing. High-profile individuals, including celebrities and politicians, are particularly vulnerable to this issue. In recent years, extensive research has been conducted to understand the creation and detection of deepfakes, with deep learning algorithms showing promising results in identifying manipulated media. This study provides an in-depth analysis of deepfake generation and detection technologies, highlighting various deep learning techniques. It also discusses the challenges posed by current detection systems and the availability of relevant databases. With the growing ease of generating and sharing deepfakes, the absence of robust detection tools is a significant global concern. This paper proposes the use of ResNext models to automatically detect deepfake videos by identifying manipulation and temporal inconsistencies between frames.
DOI Requested