Defending Deepfake via Texture Feature Perturbation

TL;DR

This paper proposes a proactive Deepfake detection method that inserts invisible texture perturbations into facial images, distorting Deepfake generation and revealing visual defects to improve detection robustness.

Contribution

It introduces a texture feature-based proactive detection framework using localized perturbations guided by attention mechanisms, a novel approach compared to passive detection methods.

Findings

Effective in distorting Deepfake generation across multiple models

Produces visible defects that aid in detection

Demonstrates promising results on CelebA-HQ and LFW datasets

Abstract

The rapid development of Deepfake technology poses severe challenges to social trust and information security. While most existing detection methods primarily rely on passive analyses, due to unresolvable high-quality Deepfake contents, proactive defense has recently emerged by inserting invisible signals in advance of image editing. In this paper, we introduce a proactive Deepfake detection approach based on facial texture features. Since human eyes are more sensitive to perturbations in smooth regions, we invisibly insert perturbations within texture regions that have low perceptual saliency, applying localized perturbations to key texture regions while minimizing unwanted noise in non-textured areas. Our texture-guided perturbation framework first extracts preliminary texture features via Local Binary Patterns (LBP), and then introduces a dual-model attention strategy to generate and optimize texture perturbations. Experiments on CelebA-HQ and LFW datasets demonstrate the promising performance of our method in distorting Deepfake generation and producing obvious visual defects under multiple attack models, providing an efficient and scalable solution for proactive Deepfake detection.