A Baseline Method for Removing Invisible Image Watermarks using Deep Image Prior

TL;DR

This paper introduces a simple black-box method using deep image prior to remove invisible watermarks from images, providing a new baseline for evaluating watermark robustness without needing watermark system details.

Contribution

It presents a novel, easy-to-implement approach for removing invisible watermarks using DIP, serving as a baseline for watermark robustness benchmarking.

Findings

DIP can reliably remove invisible watermarks from a single image.

The method preserves high image quality after watermark removal.

Limited success of black-box methods in removing training-based visible watermarks.

Abstract

Image watermarks have been considered a promising technique to help detect AI-generated content, which can be used to protect copyright or prevent fake image abuse. In this work, we present a black-box method for removing invisible image watermarks, without the need of any dataset of watermarked images or any knowledge about the watermark system. Our approach is simple to implement: given a single watermarked image, we regress it by deep image prior (DIP). We show that from the intermediate steps of DIP one can reliably find an evasion image that can remove invisible watermarks while preserving high image quality. Due to its unique working mechanism and practical effectiveness, we advocate including DIP as a baseline invasion method for benchmarking the robustness of watermarking systems. Finally, by showing the limited ability of DIP and other existing black-box methods in evading training-based visible watermarks, we discuss the positive implications on the practical use of training-based visible watermarks to prevent misinformation abuse.