An Adaptive Multi Amplitude Thresholds Embedding Algorithm for Audio Steganography

Abstract

Secure data transfer is an ever-growing challenge. Steganography refers to the art and science of secret communication. Audio steganography provides a smart solution for secure data transferring. The idea is based on hiding a secret message inside an audio signal. The challenge is to produce a stego file that can embed more data without decreasing the quality of the audio file, which may compromise the existence of the embedded message. As the technology advances, old methods become vulnerable and easy to compromise. Several methods include an encryption algorithm to add a second line of defense against a probable eavesdropper. This paper proposes a novel algorithm that aims to improve the perceptual transparency, robustness, and capacity of the embedded message. The main idea of the algorithm is to avoid embedding in the silent or near to silent intervals of the audio samples, and to scramble the embedded message as much as possible. The algorithm uses AES-256 to increase the level of robustness. The proposed algorithm is uniquely adapted to the message size and expected to perform better than the existing algorithms.