Robustness and Generalizability of Deepfake Detection: A Study with Diffusion Models

View on arXiv ← Back to list

Authors: Haixu Song, Shiyu Huang, Yinpeng Dong, Wei-Wei Tu

Published: 2023-09-05 13:22:41+00:00

AI Summary

This research introduces the DeepFakeFace (DFF) dataset, a collection of artificial celebrity faces generated using diffusion models, to improve deepfake detection. The paper proposes two evaluation methods to assess the robustness and generalizability of deepfake detection algorithms, focusing on cross-generator detection and performance on degraded images.

Abstract

The rise of deepfake images, especially of well-known personalities, poses a serious threat to the dissemination of authentic information. To tackle this, we present a thorough investigation into how deepfakes are produced and how they can be identified. The cornerstone of our research is a rich collection of artificial celebrity faces, titled DeepFakeFace (DFF). We crafted the DFF dataset using advanced diffusion models and have shared it with the community through online platforms. This data serves as a robust foundation to train and test algorithms designed to spot deepfakes. We carried out a thorough review of the DFF dataset and suggest two evaluation methods to gauge the strength and adaptability of deepfake recognition tools. The first method tests whether an algorithm trained on one type of fake images can recognize those produced by other methods. The second evaluates the algorithm's performance with imperfect images, like those that are blurry, of low quality, or compressed. Given varied results across deepfake methods and image changes, our findings stress the need for better deepfake detectors. Our DFF dataset and tests aim to boost the development of more effective tools against deepfakes.

Key findings

The results show significant variability in the performance of the RECCE deepfake detector across different generation methods and image degradation types. The study highlights the need for more robust and generalizable deepfake detection algorithms. The DFF dataset is publicly available to facilitate further research.

Approach

The research focuses on creating a new dataset, DeepFakeFace (DFF), using diffusion models to generate high-resolution deepfake images of celebrities. Two evaluation methods are proposed to test the robustness of existing deepfake detectors: cross-generator classification and degraded image classification.

Datasets

DeepFakeFace (DFF), IMDB-WIKI, FaceForensics++

Model(s)

RECCE

Author countries

China

← Previous