Publication

Abstract : Steganography is the method of concealing a message or image within a media, be it audio, image or video. The purpose is the same as that of cryptography, which is to secure data. But the difference is that in steganography, the sole existence of such a secret message is known only to the sender and receiver of that message. Audio steganography is such a technique of hiding a message within an audio file (known as the cover audio). Lately, lot of novel and versatile Audio Steganographic methods have been proposed. A good audio steganographic technique not only intends at embedding data in such a way that the changes made in the cover audio are indiscernible but also at the efficient extraction of the data. The least-significant-bit (LSB) based approach is popular among the steganographic algorithms. Our aim here is creating an audio steganography technique by changing the LSBs of the Discrete-Cosine-Transform (DCT) coefficients of the Cover Audio. It is different from the conventional LSB technique in DCT domain because, here in proposed method data is embedded by spreading it all over the layers of the LSBs of the DCT coefficients. The technique has proven to be robust than the conventional LSB technique and to the channel induced noise as well, in terms of performance evaluation. The same conclusion is drawn from the experimental results obtained by using the DCT domain features and the Support Vector Machine (SVM) as the classifier. Besides, check on the integrity of the message and imperceptibility of the changes is also performed.