A study on deep fake identification techniques using deep learning.

Autor: Lal, Kavita, Saini, Madan Lal
Předmět:
Zdroj: AIP Conference Proceedings; 2023, Vol. 2782 Issue 1, p1-12, 12p
Abstrakt: In recent years, technology for manipulating face videos has advanced to the point that a person may be unable to identify manipulation. One can insert a person into a film by pasting their face over the face of another person, or alter a person's lip movement and facial expressions to make them say whatever they want. It's easy to foresee the devastation that these kind of videos could cause if they were generated maliciously; phoney videos of politicians making outrageous statements, television presenters presenting fake news, business partners or family members requesting money, and so forth. These could have a detrimental effect on everyone, from individuals to society and democracy in general. Deep fake is a term that refers to a combination of deep learning and fake. It refers to content imitation in which the source person swaps the target subject's face to make videos or images of the target subject. Although the process of creating fake content is well-known, Deep fake goes beyond anyone's imagination, making these tactics more powerful and virtually real. Deep fake uses Machine Learning and Artificial Intelligence to change original content to create fraud one. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index