Detection of hidden object in speech based on fast fourier transform algorithm

Abstract EN

In this paper steganalysis technique isproposed on the basis of spectraldomain analysis using Discrete Fourier Transform, Fast Fourier Transform Algorithm (DFT_FFTA).

The aim from using this algorithm is to provide robust evidence for presence of hidden object in speech segment.

The Discrete Wavelet Transform (DWT) is used to decompose the speech segment, 20 seconds in length.

The speech is decomposed to the third level.

An image of 512 x 512 pixels embedded in the third level of the speech coefficients.

Reverse Discrete Wavelet Transform (RDWT) is applied to get a speech with hidden object (image) called stego-speech.

DFT_FFTA is used to analyze the stego-speech to discover an evidence of hidden object in the stegospeech. Experimental results shows that the proposed algorithm is comparable to previously existing techniques and give very clear and strong indication for the existence of stego-object.