Unmasking DeepFakes with simple Features

Authors: Ricard Durall, Margret Keuper, Franz-Josef Pfreundt, Janis Keuper

Published: 2019-11-02 09:42:25+00:00

AI Summary

This paper introduces a simple yet highly effective method for detecting DeepFake images and videos, leveraging classical frequency domain analysis. By identifying hidden artifacts in AI-generated content through power spectrum features, the approach achieves high classification accuracy. It demonstrates strong performance even with limited labeled training data and in unsupervised settings, outperforming prior systems reliant on large datasets.

Abstract

Deep generative models have recently achieved impressive results for many real-world applications, successfully generating high-resolution and diverse samples from complex datasets. Due to this improvement, fake digital contents have proliferated growing concern and spreading distrust in image content, leading to an urgent need for automated ways to detect these AI-generated fake images. Despite the fact that many face editing algorithms seem to produce realistic human faces, upon closer examination, they do exhibit artifacts in certain domains which are often hidden to the naked eye. In this work, we present a simple way to detect such fake face images - so-called DeepFakes. Our method is based on a classical frequency domain analysis followed by basic classifier. Compared to previous systems, which need to be fed with large amounts of labeled data, our approach showed very good results using only a few annotated training samples and even achieved good accuracies in fully unsupervised scenarios. For the evaluation on high resolution face images, we combined several public datasets of real and fake faces into a new benchmark: Faces-HQ. Given such high-resolution images, our approach reaches a perfect classification accuracy of 100% when it is trained on as little as 20 annotated samples. In a second experiment, in the evaluation of the medium-resolution images of the CelebA dataset, our method achieves 100% accuracy supervised and 96% in an unsupervised setting. Finally, evaluating a low-resolution video sequences of the FaceForensics++ dataset, our method achieves 91% accuracy detecting manipulated videos. Source Code: https://github.com/cc-hpc-itwm/DeepFakeDetection

Key findings

The approach achieved perfect 100% classification accuracy on high-resolution Faces-HQ and medium-resolution CelebA datasets, even when trained on as few as 20 annotated samples. For low-resolution video sequences from the FaceForensics++ dataset, the method reached 91% accuracy in detecting manipulated videos. These results demonstrate the robustness and effectiveness of frequency domain analysis for deepfake detection across various resolutions and with minimal data requirements.

Approach

The method involves transforming input images to grayscale and applying Discrete Fourier Transform (DFT) to analyze their frequency components. An azimuthal average is then computed from the power spectrum to create robust 1D features. These frequency-domain features are subsequently fed into basic classifiers such as Support Vector Machines (SVMs), Logistic Regression, or K-Means clustering for fake detection.

Datasets

Faces-HQ (created by authors, combining CelebA-HQ, Flickr-Faces-HQ, 100K Faces project, www.thispersondoesnotexist.com), CelebA, FaceForensics++ (DeepFakeDetection Data set).

Model(s)

UNKNOWN

Author countries

Germany

← Previous