Adaptive Wavelet Domain Audio Steganography with High Capacity and Low Error Rate

Abstract

Embedding a secret message into a cover media without attracting attention, which is known as steganography, is desirable in some security applications. One of the medias which can be used as a cover media is audio signal. In this paper we introduce an adaptive wavelet domain steganography with high capacity and low error rate. We use lifting scheme to create perfect reconstruction filter banks which are Int2Int and hide data in least significant bits (LSB) of details coefficient in an adaptive way to reduce the error rate. Our method have zero error rate for hiding capacity below 100 kilo bits-per-second (kbps) and 0.3% error for 200 kbps, in comparison to 0.9% error of normal wavelet domain LSB steganography. Signal to noise ratio (SNR) values and listening tests results show that the stegano audio is imperceptible from original audio even with hiding capacity up to 200 kbps.