From Pixels to Prompts: A Systematic Study and Introduction to Image Prompt Injection Attacks

View at publisher Summarise with VIRAT
Authors: Neha Nagaraja, Lan Zhang, Zhilong Wang
Cited by: 0
References: 21

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This article introduces and organizes image prompt injection as a practical threat where visually subtle text in natural images can steer multimodal large language models.

Fields of Study & Subjects

Adversarial Robustness in Machine Learning Digital Media Forensic Detection Advanced Steganography and Watermarking Techniques

Publication details

Journal / source
Computer
Volume
59
Issue
4
Pages
105–110
Publication date
31 March 2026
Language
EN