Realism to Deception: Investigating Deepfake Detectors Against Face Enhancement

View on arXiv ← Back to list

Authors: Muhammad Saad Saeed, Ijaz Ul Haq, Khalid Malik

Published: 2025-09-08 19:49:52+00:00

AI Summary

This research investigates how face enhancement techniques, both traditional and GAN-based, affect the accuracy of deepfake detectors. The study finds that these enhancement methods significantly reduce the accuracy of deepfake detection, acting as an anti-forensic tool by masking artifacts used for detection.

Abstract

Face enhancement techniques are widely used to enhance facial appearance. However, they can inadvertently distort biometric features, leading to significant decrease in the accuracy of deepfake detectors. This study hypothesizes that these techniques, while improving perceptual quality, can degrade the performance of deepfake detectors. To investigate this, we systematically evaluate whether commonly used face enhancement methods can serve an anti-forensic role by reducing detection accuracy. We use both traditional image processing methods and advanced GAN-based enhancements to evaluate the robustness of deepfake detectors. We provide a comprehensive analysis of the effectiveness of these enhancement techniques, focusing on their impact on Na\\ive, Spatial, and Frequency-based detection methods. Furthermore, we conduct adversarial training experiments to assess whether exposure to face enhancement transformations improves model robustness. Experiments conducted on the FaceForensics++, DeepFakeDetection, and CelebDF-v2 datasets indicate that even basic enhancement filters can significantly reduce detection accuracy achieving ASR up to 64.63\\%. In contrast, GAN-based techniques further exploit these vulnerabilities, achieving ASR up to 75.12\\%. Our results demonstrate that face enhancement methods can effectively function as anti-forensic tools, emphasizing the need for more resilient and adaptive forensic methods.

Key findings

Face enhancement methods, especially GAN-based ones, significantly reduced deepfake detection accuracy (ASR up to 75.12%). Even basic enhancement filters caused substantial accuracy drops (ASR up to 64.63%). The results highlight a trade-off between improved perceptual quality and reduced detection accuracy.

Approach

The researchers evaluated the robustness of six deepfake detectors (two each of Naïve, Spatial, and Frequency-based) against six face enhancement methods (including Gaussian smoothing, bilateral filtering, and GAN-based methods) applied to three datasets. They measured attack success rate (ASR) and perceptual quality to assess the effectiveness of the face enhancement as anti-forensic tools.

Datasets

FaceForensics++, DeepFakeDetection, CelebDF-v2

Model(s)

EfficientNet-B4, Xception, CORE, UCF, F3Net, SPSL

Author countries

USA

← Previous