Towards quantum audio steganalysis using synergy of quantum fourier transform and quantum neural network

Abstract

The proposed approach in this study introduces a comprehensive audio steganalysis scheme that integrates quantum signal processing with machine learning techniques. This method employs the quantum Fourier transform on the Quantum Representation of Digital Signals (QRDS) to extract statistical features from the second-order derivatives of the audio spectrum. These features are derived by analyzing the rate of change in the gradient of the quantum spectrum, providing valuable insights for identifying steganographic content, concealed within the audio data. The statistical analysis of these features includes the quantum spectral center (QSC), quantum spectral bandwidth (QSB), quantum spectral flatness measurement (QSFM), and quantum spectral crest factor (QSFC). The extracted features are then input into a multilayer quantum neural network that utilizes simple quantum gates, thereby reducing the algorithm's complexity and the time required for training and testing. The classification algorithm, applied by this neural network, can distinguish between clean and stego audio datasets, with an accuracy exceeding 96 %. It outperforms existing methods in both efficiency and accuracy.