IDEAW: Robust Neural Audio Watermarking with Invertible Dual-Embedding

TL;DR

IDEAW introduces a robust neural audio watermarking model with dual-embedding for better capacity and locating, addressing limitations of previous neural methods through invertible neural networks and attack layer considerations.

Contribution

We propose a novel dual-embedding neural watermarking model, IDEAW, that improves robustness, capacity, and locating ability by integrating invertible neural networks and attack layer effects.

Findings

Outperforms existing methods in robustness against attacks

Achieves higher watermark capacity

Demonstrates efficient watermark locating

Abstract

The audio watermarking technique embeds messages into audio and accurately extracts messages from the watermarked audio. Traditional methods develop algorithms based on expert experience to embed watermarks into the time-domain or transform-domain of signals. With the development of deep neural networks, deep learning-based neural audio watermarking has emerged. Compared to traditional algorithms, neural audio watermarking achieves better robustness by considering various attacks during training. However, current neural watermarking methods suffer from low capacity and unsatisfactory imperceptibility. Additionally, the issue of watermark locating, which is extremely important and even more pronounced in neural audio watermarking, has not been adequately studied. In this paper, we design a dual-embedding watermarking model for efficient locating. We also consider the impact of the attack layer on the invertible neural network in robustness training, improving the model to enhance both its reasonableness and stability. Experiments show that the proposed model, IDEAW, can withstand various attacks with higher capacity and more efficient locating ability compared to existing methods.