Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deepfake Detection: A Comprehensive Survey from the Reliability Perspective

Tianyi Wang, Xin Liao, Kam Pui Chow, Xiaodong Lin, Yinglong Wang

TL;DR

This survey provides a thorough review of the existing Deepfake detection studies from the reliability perspective using statistical random sampling knowledge and the publicly available benchmark datasets to review the reliability of the existing detection models on arbitrary Deepfake candidate suspects.

Abstract

The mushroomed Deepfake synthetic materials circulated on the internet have raised a profound social impact on politicians, celebrities, and individuals worldwide. In this survey, we provide a thorough review of the existing Deepfake detection studies from the reliability perspective. We identify three reliability-oriented research challenges in the current Deepfake detection domain: transferability, interpretability, and robustness. Moreover, while solutions have been frequently addressed regarding the three challenges, the general reliability of a detection model has been barely considered, leading to the lack of reliable evidence in real-life usages and even for prosecutions on Deepfake-related cases in court. We, therefore, introduce a model reliability study metric using statistical random sampling knowledge and the publicly available benchmark datasets to review the reliability of the existing detection models on arbitrary Deepfake candidate suspects. Case studies are further executed to justify the real-life Deepfake cases including different groups of victims with the help of the reliably qualified detection models as reviewed in this survey. Reviews and experiments on the existing approaches provide informative discussions and future research directions for Deepfake detection.

Deepfake Detection: A Comprehensive Survey from the Reliability Perspective

TL;DR

This survey provides a thorough review of the existing Deepfake detection studies from the reliability perspective using statistical random sampling knowledge and the publicly available benchmark datasets to review the reliability of the existing detection models on arbitrary Deepfake candidate suspects.

Abstract

The mushroomed Deepfake synthetic materials circulated on the internet have raised a profound social impact on politicians, celebrities, and individuals worldwide. In this survey, we provide a thorough review of the existing Deepfake detection studies from the reliability perspective. We identify three reliability-oriented research challenges in the current Deepfake detection domain: transferability, interpretability, and robustness. Moreover, while solutions have been frequently addressed regarding the three challenges, the general reliability of a detection model has been barely considered, leading to the lack of reliable evidence in real-life usages and even for prosecutions on Deepfake-related cases in court. We, therefore, introduce a model reliability study metric using statistical random sampling knowledge and the publicly available benchmark datasets to review the reliability of the existing detection models on arbitrary Deepfake candidate suspects. Case studies are further executed to justify the real-life Deepfake cases including different groups of victims with the help of the reliably qualified detection models as reviewed in this survey. Reviews and experiments on the existing approaches provide informative discussions and future research directions for Deepfake detection.
Paper Structure (29 sections, 5 equations, 2 figures, 8 tables, 1 algorithm)

This paper contains 29 sections, 5 equations, 2 figures, 8 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. The Evolution of Deepfake Generation
  3. Deepfake Generation
  4. Deepfake Benchmark Datasets
  5. Reliability-Oriented Challenges of Deepfake Detection
  6. Transferability
  7. Interpretability
  8. Robustness
  9. Detection Model Reliability Study
  10. Overview
  11. Deepfake Population
  12. Random Sampling
  13. Dataset Preparation
  14. Dataset Pre-processing
  15. Datasets Involved and Detailed Arrangements
  16. ...and 14 more sections

Figures (2)

  • Figure 1: Demonstrations of the three challenges from top to bottom. Transferability (top) refers to models that focus on stable detection ability on unseen benchmark datasets; interpretability (middle) refers to efforts on explaining the model detected falsification; robustness (bottom) refers to models that handle Deepfake suspects under different real-life conditions and scenarios.
  • Figure 2: Screenshots of the four real-life Deepfake videos for Deepfake detection in the case study. The videos are hyper-realistic with different resolutions and no obvious artifacts can be observed visually by human eyes.