Deep Robust Reversible Watermarking

2025-03-04

AI Summary

Traditional robust reversible watermarking (RRW) methods suffer from complex design, high computational overhead, low recovery fidelity, and weak robustness. To address these issues, this paper proposes the first end-to-end trainable deep integer-reversible watermarking framework. Its core contributions are: (1) an integer-reversible watermarking network (iIWN) featuring an encoder–noise layer–decoder architecture integrated with arithmetic coding to guarantee strict reversibility; (2) an overflow penalty loss and auxiliary weight adaptation strategy that eliminates distortion-specific handcrafting and enables adaptive robustness; and (3) lossless dual recovery of both host and watermark, with stable extraction even over lossy channels. Experiments show that embedding, extraction, and recovery complexities are reduced by 55.14×, 5.95×, and 3.57×, respectively; auxiliary bitstream compression achieves a 43.86× rate; 100% reversibility is achieved on 16,762 PASCAL VOC images; and the method surpasses state-of-the-art irreversible approaches in both robustness and visual quality.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Robust Reversible Watermarking (RRW) enables perfect recovery of cover images and watermarks in lossless channels while ensuring robust watermark extraction in lossy channels. Existing RRW methods, mostly non-deep learning-based, face complex designs, high computational costs, and poor robustness, limiting their practical use. This paper proposes Deep Robust Reversible Watermarking (DRRW), a deep learning-based RRW scheme. DRRW uses an Integer Invertible Watermark Network (iIWN) to map integer data distributions invertibly, addressing conventional RRW limitations. Unlike traditional RRW, which needs distortion-specific designs, DRRW employs an encoder-noise layer-decoder framework for adaptive robustness via end-to-end training. In inference, cover image and watermark map to an overflowed stego image and latent variables, compressed by arithmetic coding into a bitstream embedded via reversible data hiding for lossless recovery. We introduce an overflow penalty loss to reduce pixel overflow, shortening the auxiliary bitstream while enhancing robustness and stego image quality. An adaptive weight adjustment strategy avoids manual watermark loss weighting, improving training stability and performance. Experiments show DRRW outperforms state-of-the-art RRW methods, boosting robustness and cutting embedding, extraction, and recovery complexities by 55.14( imes), 5.95( imes), and 3.57( imes), respectively. The auxiliary bitstream shrinks by 43.86( imes), with reversible embedding succeeding on 16,762 PASCAL VOC 2012 images, advancing practical RRW. DRRW exceeds irreversible robust watermarking in robustness and quality while maintaining reversibility.

Problem

• Enhances robustness and quality of reversible watermarking.
• Reduces computational costs and complexity in watermarking.
• Improves training stability and performance via adaptive strategies.

Innovation

• Deep learning-based robust reversible watermarking scheme
• Integer Invertible Watermark Network for invertible mapping
• Encoder-noise layer-decoder framework for adaptive robustness