FreqMark: Invisible Image Watermarking via Frequency Based Optimization in Latent Space

TL;DR

FreqMark is a novel frequency-based latent space optimization method for invisible image watermarking that enhances robustness against regeneration attacks while maintaining high image quality.

Contribution

It introduces a new watermarking approach using latent frequency space optimization post-VAE encoding, improving robustness and flexibility over existing methods.

Findings

Achieves over 90% bit accuracy in watermark extraction.

Demonstrates superior robustness against regeneration attacks.

Maintains high image quality with flexible encoding options.

Abstract

Invisible watermarking is essential for safeguarding digital content, enabling copyright protection and content authentication. However, existing watermarking methods fall short in robustness against regeneration attacks. In this paper, we propose a novel method called FreqMark that involves unconstrained optimization of the image latent frequency space obtained after VAE encoding. Specifically, FreqMark embeds the watermark by optimizing the latent frequency space of the images and then extracts the watermark through a pre-trained image encoder. This optimization allows a flexible trade-off between image quality with watermark robustness and effectively resists regeneration attacks. Experimental results demonstrate that FreqMark offers significant advantages in image quality and robustness, permits flexible selection of the encoding bit number, and achieves a bit accuracy exceeding 90% when encoding a 48-bit hidden message under various attack scenarios.