Deepfake Forensics: Identifying Real Regions in Altered Videos with Digital Watermarking
Year: 2023
Type: article
Abstract: This paper describes a method for detecting Deepfake videos using a lightweight yet secure video encryption algorithm. With the increasing use of digital media, transferring data via the Internet or other mediums requires protection. In the proposed method a digital signature is generated and encrypted using Asymmetric Encryption (RSA). This encrypted signature is then used as a blind watermark for the video. This technique aims to detect “face swap” type of Deepfake videos. It is an efficient algorithm and has minimal impact on the perceptibility of the video quality.
Source: Journal of Student Research
Authors: Aayush Asthana, Sam Saarinen
Institution: Brown University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4390735322
