Non-Target Conversion Based Speech Steganography for Secure Speech Communication System
Mingjun Zhang, Yan Feng, Yu Miao Gao, Longting Xu · 2024
The widespread application of speech data increases the risk of speaker identity being compromised during speech communication. To mitigate this risk and protect voice privacy, we propose a system aimed at ensuring the security of speech communication by avoiding the exposure of the actual speaker’s identity. Our system comprises three main components. Firstly, the non-target voice conversion system based on generative adversarial network converts the original audio into the audio of a non-existent person while preserving the speaker embedding of the real audio. Secondly, during speech communication, we utilize speech steganography techniques to embed the actual speaker embedding into the converted audio. Finally, at the receiving end, we extract the actual speaker embedding from the transmitted converted audio and use it to reconstruct the original audio. Experimental results validate the effectiveness of our system, showcasing an innovative solution in the field of speech security.