Audio Steganography Method, Device, and Electronic Equipment Based on Quantum Generative Adversarial Networks

Abstract

This application discloses an audio steganography method, apparatus, and electronic device based on a quantum generative adversarial network (GAN). The method, applied at a transmitting end, includes: embedding the information to be steganized into original quantum initial-state data to obtain dense quantum initial-state data; performing time-domain feature processing on the dense quantum initial-state data through a generative network and obtaining the corresponding frequency-domain quantum state data; performing frequency-domain refinement processing on the frequency-domain quantum state data; converting the obtained frequency-domain refined data into target quantum state audio data through an inverse quantum Fourier transform; transmitting the target quantum state audio data to a receiving end through a quantum channel; and transmitting the quantum circuit parameters corresponding to the generative network to the receiving end through a classical channel. Implementing this application embodiment can optimize the time-frequency mixing processing flow while ensuring secure audio data transmission, effectively improving the imperceptibility of the steganographic information, thereby enhancing the security of information hiding.

Images