A Novel Steganalysis of Steghide Focused on High-Frequency Region of Audio Waveform

In this study, steganalysis of steghide embedded in Microsoft RIFF waveform audio format (WAV) data is investigated. Spectral analyses show that the conventional steganalysis utilize the statistics of high-frequency regions and silent temporal segments of the target signals intentionally or unintentionally. Moreover, the frequency components just below the Nyquist frequency are important for statistic-based steganalysis in terms of the signal-to-noise ratio where the signal corresponds to the distortion components induced by data hiding, and the noise corresponds to the cover signal. A novel steganalysis making full use of the high- frequency features is proposed, and its detection performance is compared with the conventional method, which showed the best performance so far. The results show that the proposed steganalysis outperforms the conventional method for cover data of 100 music signals and 320 speech signals mixed with background noises.