Article Preview
Top1. Introduction
Steganography is an art of hiding secret information in another seemingly innocuous message, or carrier (Johnson, Duric&Jajodia, 2001). Steganography is different from cryptography. The primary goal of cryptography is to provide secure communication by changing data into a form that it cannot understand except for the sender and the receiver. Since when an attacker does not know the existence of secret message, he will not generate the idea of attacking it. Therefore, steganography is a powerful technique to enhance the security of data transmission. Throughout history, a multitude of methods have been used to hide information. With the development of the Internet and other new technologies, digital steganography technique which is used to embed the secret message into digital multimedia is gradually rising. It has developed a strong basis for the area of steganography with a growing number of applications for digital fields like covert communications, annotation etc. So far, various researches on steganography have been carried out on storage media, such as text, image, audio, and video.
Furthermore, auditory system is the largest source of information in addition to the Human Visual System(HVS). Thus the research of audio steganography is of great significance and has wild application scenarios.
Like steganography in other media, audio steganographic technique also has three evaluation metrics (Divya&Reddy, 2012):
- 1.
Capacity means the amount of secret information that can be embedded into the host audio without affecting the perceptual quality of audio.
- 2.
Imperceptibility evaluates how well a secret message is embedded into the cover audio. The difference between audio after hiding and audio before hiding should remain negligible.
- 3.
Robustness indicates the ability of secret message to resist against attacks.
In audio steganography, Human Auditory System (HAS) is used to hide information in the audio. Because HAS has more precision than HVS, audio steganography is more challenging than image steganography (Shirali-Shahreza&Manzuri-Shalmani, (2007).
In this paper, the authors propose an audio steganography method based on run length encoding and integer wavelet transform. The ideal of this method is to design the embedding and extracting procedure of the secret information, and then the authors test the capacity, robustness and anti-detection and compare it with some existing methods. The advantages of the proposed method is large capacity.
The rest of the paper is organized as follows: the related work is introduced in Section II. In Section III the proposed methods are presented. Experimental results and analyses are presented in Section IV. Finally, Section V concludes this paper.
TopSteganography techniques can be classified into methods in time and transform domain. The common used methods of steganography in time domain include Least Significant Bits (LSB) (Cvejic&Seppänen, 2004; Roy, Parida, Singh &Sairam, 2012; Vimal& Alex, 2014), Echo Hiding (Chen& Wu, 2008; Yan, Sun & Lu, 2003), Spread Spectrum Method (Rupanshi, Preeti&Vandana, 2014) and etc. Besides, methods in transform domain contain hiding method based on DFT, DCT (Chen, Zhang, Liu, Niu, & Yang,2009; Premalatha, Narayanan,Vikash& Ramesh,2014; Chilhate, Patidar&Chandel, 2015) and DWT (Sheikhan, Asadollahi&Shahnazi, 2011; Bhattacharyya &Sanyal, 2012). It is hard to balance the capacity, imperceptibility and robustness. For instance, LSB has great hiding capacity, however, which is at the expense of more robust. Although methods in transform domain boost up robustness against some attacks, the capacity and imperceptibility are influenced a little.
Vimal and Alex (2014) discussed in their research that dual randomness LSB method was proposed which hides the secret message in randomly selected samples and bits. This method provided more confidentiality compared with conventional LSB method. But compared with the method of Cvejic and Seppanen (2003), the hiding capacity was smaller. Cvejic and Seppanen (2003) showed in their work that a novel modification to standard LSB algorithm was proposed, which was able to embed four bits per sample and improved the capacity of data hiding channel by 33%. They all belong to LSB method whose disadvantage is weak robust.