Home / Papers / Deepfake Creation Using Gans and Autoencoder and Deepfake detection

Deepfake Creation Using Gans and Autoencoder and Deepfake detection

2 Citations2023
M. Das, Manav Kumar, Ishank Kumar Kapil
2023 2nd International Conference on Vision Towards Emerging Trends in Communication and Networking Technologies (ViTECoN)

In a fake video, the face, expression or speech is replaced with an image of another's face, with a distinct speech or emotion with the help of the technology of deep learning.

Abstract

Deep learning techniques is useful in a myriad of applications such as computer vision processing, natural language processing, as well as deepfake detection. The development of deep learning algorithms for imaging detection has resulted in the development of deepfakes. These fakes employ advanced algorithms for deep learning to generate fake images that are extremely difficult to differentiate from authentic images. In a fake video, the face, expression or speech is replaced with an image of another's face, with a distinct speech or emotion with the help of the technology of deep learning. These videos are typically so sophisticated that the traces of manipulation are hard to spot. Social media are among the most frequently targeted and most serious because they are vulnerable platforms that are susceptible to blackmailing or making a person look bad. There are several existing efforts to detect fake images, but there have been very few efforts developed for video content on social media.