Covert speech communication through audio steganography using DWT and SVD

In this paper, a covert speech communication technique through a semi-blind audio steganography algorithm using discrete wavelet transform (DWT) and singular value decomposition (SVD) is proposed. This method securely transmits the speech signal by embedding its singular values (SV) in a cover audio. The cover audio is decomposed using fifth level DWT and embedding is performed in SV matrix of approximate coefficients to achieve good imperceptibility and robustness to the attacks. The performance of the proposed algorithm is evaluated by conducting imperceptibility and robustness tests on various speech signals. The experimental results show that the secret speech is reconstructed with high SNR, correlation and Perceptual Evaluation of Speech Quality (PESQ) score even the stego audio is subjected to various attacks. It is observed that under no attack conditions the correlation coefficient (NCC) is 1.0, SNR and PESQ of reconstructed secret audio is 70.58 and 4.49, respectively.