Protective Perturbations against Unauthorized Data Usage in Diffusion-based Image Generation

TL;DR

This paper surveys protective perturbation techniques that safeguard against unauthorized data usage in diffusion-based image generation, addressing privacy and intellectual property concerns in AI models.

Contribution

It provides a comprehensive survey, establishes a threat model, categorizes tasks, and proposes an evaluation framework for protective perturbation methods.

Findings

Categorizes existing protective perturbation methods

Analyzes designs and effectiveness of perturbation techniques

Proposes a standardized evaluation framework

Abstract

Diffusion-based text-to-image models have shown immense potential for various image-related tasks. However, despite their prominence and popularity, customizing these models using unauthorized data also brings serious privacy and intellectual property issues. Existing methods introduce protective perturbations based on adversarial attacks, which are applied to the customization samples. In this systematization of knowledge, we present a comprehensive survey of protective perturbation methods designed to prevent unauthorized data usage in diffusion-based image generation. We establish the threat model and categorize the downstream tasks relevant to these methods, providing a detailed analysis of their designs. We also propose a completed evaluation framework for these perturbation techniques, aiming to advance research in this field.