Robust Audio Data Hiding based Probabilistic Global Search in MDCT Domain

Abstract

This paper managed to propose robust audio data hiding through optimization-based problem transformation method, Probabilistic Global Search Lausanne Algorithm (PGSL) in modified discrete cosine transform (MDCT) domain. We manipulate the MDCT coefficients in the selected frames with the stego key and, then implement the optimal best points searching process through PGSL Algorithm to obtain the best secret audio message embedding positions. The Lowest Bit Error Rate value are defined the optimal best coefficients. The inverse modified discrete cosine transform (IMDCT) is applied to form stego audio. In our audio data hiding scheme, the results of imperceptible (audible difference) between the stego and cover audio will be evaluated using Spectrogram and Informal Listening Tests. The robustness will be tested with standard digital signal manipulation attacks as Noise addition, MPEG compression and Resampling. The goal of proposed data hiding is to protect the privacy of secret audio message from unauthorized accesses, and to satisfy audio steganography issues as imperceptible, embedding capacity and robustness