Deepfake: definitions, performance metrics and standards, datasets, and a meta-review

2024/09/04 by Enes Altuncu, Virginia N. L. Franqueira, Shujun Li · 1 voice · 1 citation
Computer Science · #Digital Media Forensic Detection #Generative Adversarial Networks and Image Synthesis #Image and Video Quality Assessment

paper · pdf · doi:10.3389/fdata.2024.1400024

openalex publication_date 2024/09/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/04

Abstract

Recent advancements in AI, especially deep learning, have contributed to a significant increase in the creation of new realistic-looking synthetic media (video, image, and audio) and manipulation of existing media, which has led to the creation of the new term "deepfake." Based on both the research literature and resources in English, this paper gives a comprehensive overview of deepfake, covering multiple important aspects of this emerging concept, including (1) different definitions, (2) commonly used performance metrics and standards, and (3) deepfake-related datasets. In addition, the paper also reports a meta-review of 15 selected deepfake-related survey papers published since 2020, focusing not only on the mentioned aspects but also on the analysis of key challenges and recommendations. We believe that this paper is the most comprehensive review of deepfake in terms of the aspects covered.

Citations

• WaveNet: A Generative Model for Raw Audio
• FaceForensics: A Large-scale Video Dataset for Forgery Detection in Human Faces
• WaveCycleGAN2: Time-domain Neural Post-filter for Speech Waveform Generation
• ASVspoof 2019: A large-scale public database of synthesized, converted and replayed speech
• EmoFake: An Initial Dataset for Emotion Fake Audio Detection
• How Close is ChatGPT to Human Experts? Comparison Corpus, Evaluation, and Detection
• MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models
• X-IQE: eXplainable Image Quality Evaluation for Text-to-Image Generation with Visual Large Language Models
• Robustness and Generalizability of Deepfake Detection: A Study with Diffusion Models
• AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset
• Towards A Better Metric for Text-to-Video Generation
• MLAAD: The Multi-Language Audio Anti-Spoofing Dataset
• Unmasking DeepFakes with simple Features

Cited by

• Ensemble Deep Learning Approaches for AI-Altered Video Detection

Discussions

• Deepfake is a portmanteau of "deep learning" and "fake," referring to the use of deep learning algorithms to generate fake multimedia (images, videos, audio, and even text!). See our recent review fo [bsky, 0 points, 1 comments]