An Automated Workflow for Deepfake Detection

EasyChair Preprint 11044

Abstract

The lightning development of Artificial Intelligence (AI) has brought significant changes to modern society, including the emergence of AI-generated art and image enhancement techniques. However, one of the most alarming consequences of AI advancement is the creation of deepfake images and videos through the use of General Adversarial Networks (GANs). As these deepfakes become increasingly convincing and widespread, there is an urgent need for measures to detect and prevent their dissemination, especially on independent social-media websites. The proposed approach results in accuracy scores comparable to and surpassing several SOTA(State-of-the-art) approaches on three benchmark datasets, while consuming considerably lesser computational overhead, and containing over 100x lesser trainable parameters which was achieved using the extraction and manipulation of geometrical features.

Keyphrases: Autonmation, Bi-LSTM, Celeb-DF, Deep Fake Detection, FF++, Lucas-Kande, UADFV, computer vision, facial landmarks

@booklet{EasyChair:11044,
  author    = {Anirudh Joshi and Chandrashekar Chavan},
  title     = {An Automated Workflow for Deepfake Detection},
  howpublished = {EasyChair Preprint 11044},
  year      = {EasyChair, 2023}}