WavInWav: Time-domain Speech Hiding via Invertible Neural Network

2025-10-03

AI Summary

Existing audio steganography methods suffer from poor secret speech reconstruction quality and weak robustness due to insufficient time-frequency modeling capability. To address this, we propose WavInWav—the first end-to-end, purely time-domain reversible speech steganography framework. It employs a flow-based invertible neural network to establish bidirectional lossless mappings among cover, stego, and secret speech signals; introduces a time-frequency joint loss function that explicitly enforces spectral consistency during time-domain optimization; and integrates a lightweight encryption mechanism to ensure semantic security. Evaluated on VCTK and LibriSpeech, WavInWav achieves state-of-the-art performance across objective (PESQ, STOI) and subjective (MOS) metrics. It demonstrates strong robustness against common distortions—including additive noise and channel impairments—and supports real-time embedding and extraction. The framework is thus well-suited for high-fidelity, high-security speech steganographic communication.

Technology Category

• Natural Language Processing: Speech
• Machine Learning: Privacy
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Security and Privacy: Applications of cryptography
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Data hiding is essential for secure communication across digital media, and recent advances in Deep Neural Networks (DNNs) provide enhanced methods for embedding secret information effectively. However, previous audio hiding methods often result in unsatisfactory quality when recovering secret audio, due to their inherent limitations in the modeling of time-frequency relationships. In this paper, we explore these limitations and introduce a new DNN-based approach. We use a flow-based invertible neural network to establish a direct link between stego audio, cover audio, and secret audio, enhancing the reversibility of embedding and extracting messages. To address common issues from time-frequency transformations that degrade secret audio quality during recovery, we implement a time-frequency loss on the time-domain signal. This approach not only retains the benefits of time-frequency constraints but also enhances the reversibility of message recovery, which is vital for practical applications. We also add an encryption technique to protect the hidden data from unauthorized access. Experimental results on the VCTK and LibriSpeech datasets demonstrate that our method outperforms previous approaches in terms of subjective and objective metrics and exhibits robustness to various types of noise, suggesting its utility in targeted secure communication scenarios.

Problem

• Enhancing secret audio recovery quality in data hiding
• Addressing time-frequency transformation limitations in audio steganography
• Improving reversibility and security for secure communication systems

Innovation

• Flow-based invertible neural network for audio embedding
• Time-frequency loss applied to time-domain signals
• Encryption technique to protect hidden data