Logo image
Back
Beyond Deepfake Images: Detecting AI-Generated Videos
Conference proceeding   Open access

Beyond Deepfake Images: Detecting AI-Generated Videos

Danial Samadi Vahdati, Tai D. Nguyen, Aref Azizpour and Matthew C. Stamm
2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp 4397-4408
17 Jun 2024
DOI: https://doi.org/10.1109/CVPRW63382.2024.00443
url
https://arxiv.org/abs/2404.15955View
Submitted Open arXiv.org - Non-exclusive license to distribute

Abstract

Conferences Deep Learning Deep Neural Networks Deepfake Deepfakes Detectors Forensics Generative AI Generators image forensics Media Forensics synthetic image detection Synthetic images synthetic video detection Synthetic Videos Training video forensics
Recent advances in generative AI have led to the development of techniques to generate visually realistic synthetic video. While a number of techniques have been developed to detect AI-generated synthetic images, in this paper we show that synthetic image detectors are unable to detect synthetic videos. We demonstrate that this is because synthetic video generators introduce substantially different traces than those left by image generators. Despite this, we show that synthetic video traces can be learned, and used to perform reliable synthetic video detection or generator source attribution even after H.264 re-compression. Furthermore, we demonstrate that while detecting videos from new generators through zero-shot transferability is challenging, accurate detection of videos from a new generator can be achieved through few-shot learning.

Metrics

29 Record Views
8 citations in Scopus

Details

Logo image