Not Just Text: Uncovering Vision Modality Typographic Threats in Image Generation Models

TL;DR

This paper uncovers vulnerabilities in image generation models related to the vision modality, demonstrating susceptibility to typographic attacks and evaluating defense methods, while introducing a new dataset for assessing such threats.

Contribution

It reveals that image generation models are vulnerable to vision modality threats, introduces the VMT-IGMs dataset, and evaluates existing defenses' effectiveness against these threats.

Findings

Image models are susceptible to typographic attacks in the vision modality.

Existing defense methods are largely ineffective against vision modality threats.

The VMT-IGMs dataset provides a baseline for future vulnerability assessments.

Abstract

Current image generation models can effortlessly produce high-quality, highly realistic images, but this also increases the risk of misuse. In various Text-to-Image or Image-to-Image tasks, attackers can generate a series of images containing inappropriate content by simply editing the language modality input. To mitigate this security concern, numerous guarding or defensive strategies have been proposed, with a particular emphasis on safeguarding language modality. However, in practical applications, threats in the vision modality, particularly in tasks involving the editing of real-world images, present heightened security risks as they can easily infringe upon the rights of the image owner. Therefore, this paper employs a method named typographic attack to reveal that various image generation models are also susceptible to threats within the vision modality. Furthermore, we also evaluate the defense performance of various existing methods when facing threats in the vision modality and uncover their ineffectiveness. Finally, we propose the Vision Modal Threats in Image Generation Models (VMT-IGMs) dataset, which would serve as a baseline for evaluating the vision modality vulnerability of various image generation models.