International Journal of Computer Applications (0975 – 8887) Volume 59– No.17, December 2012 1 A Novel Adaptive Steganographic Technique using Kohonen Neural Network based on Integer Wavelet Transform Liya Lesly Department of Electronics and communication Federal Institute of Science and Technology (FISAT) Rajesh Cherian Roy Department of Electronics and communication Federal Institute of Science and Technology (FISAT) ABSTRACT Steganography allows us to hide a desired confidential information in a covered or host image while image encryption is decomposing the desired information to a non- readable, non-comprehended manner. Steganography gained importance in the past few years due to the increasing need for providing secrecy in an open environment like the internet. This is made possible by hiding the very existence of the message and make communication undetectable. Steganography is the science of secret message delivery using cover media and has many technical challenges such as high hiding capacity and security. In this paper, a method is proposed to optimize these two requirements by suggesting a novel technique for hiding data in digital images. For this, an adaptive hiding capacity function along with a Kohonen neural network has been employed. This method hides secret data in Integer wavelet coefficients. The Optimal pixel adjustment (OPA) algorithm is applied after embedding the message. A new high capacity image steganographic method based on Kohonen neural network is also introduced. Kohonen network is trained according to the absolute contrast sensitivity of pixels present in cover image. On the receiving side, the original image is not needed for extracting the embedded data. It is observed that the capacity and security is increased with acceptable PSNR and hiding capacity. Keywords Steganography, Integer Wavelet Transform, Discrete Cosine Transform, adaptive algorithm, Kohonen NN, Optimal Pixel Adjustment Algorithm. 1. INTRODUCTION Information hiding is a method of thrashing secret data into a host medium so that the hidden data are imperceptible but known to the intended recipient [1]. This can be achieved by concealing the existence of information within seemingly harmless carrier or cover. Steganography means “covered writing” and it [2] concerns itself with ways of embedding a secret message into a cover object, without altering the properties of the cover object evidently. The embedding procedure is typically related with a key, usually a stego-key. Without knowledge of this key it will be difficult for a third party to extract the message or even detect its existence. Once the cover object has data embedded in it, it is called stego object. Steganography differs from cryptography in the sense that it tries to hide the message instead of transforming it so as to obscure its meaning [3].The host medium may be a digital image, audio, video or any other type of media. Among the different kinds of media, the digital image is most popularly used as the host media to convey secret information. This is due to the fact that they are easy to obtain with reasonable hiding capacity and distortion tolerance [4]. The data hiding schemes can be categorized in two groups: spatial and frequency domain. The first technique is based on embedding message in the Least Significant Bits (LSB) of image pixels. The basic LSB method has a simple implementation and high hiding capacity [5]. However, it has low robustness against some attacks such as low-pass filtering and compression [6]. An alternate method of LSB substitution is OPAP (Optimal Pixel Adjustment Process) [4] in which the image quality of the stego-image can be enhanced sufficiently. In the second method, frequency coefficients of images are selected and then embeds the messages within them. Different transforms are available for data hiding, such as Discrete Fourier Transform (DFT), the Discrete Wavelet Transform (DWT), Integer Wavelet Transform (IWT), and the Discrete Cosine Transform (DCT). JPEG, a standard image compression technique employs DCT. An example of utilizing DWT is the employment of an adaptive data embedding technique by using OPAP in which the secret data is hidden in the Integer Wavelet coefficients of the cover image [7]. This will maximize the hiding capacity as much as possible by overcoming the problems of complex computations and distortions. A pseudo-random generator function is used to select different embedding positions along with a Kohonen neural network, in which pixels are classified into different classes of sensitivity [8]. Kohonen neural networks are used because they are a relatively simple network to construct that can be trained very rapidly. Thus the adaptive data hiding function along with Kohonen network increases the system security. The rest of the paper is organized as follows: Section II provides a brief prefatory to Integer wavelet transform and Kohonen neural network. In Section III the proposed system is presented. Experimental results are illustrated in Section IV, prior to Conclusions in Section V. 2. THE STEGANOGRAPHY METHOD 2.1 Integer Wavelet Transform (IWT) In recent years, there has been a growing interest in Integer Wavelet Transforms for image processing applications. Such transforms are invertible in finite-precision arithmetic (i.e., reversible), map integers to integers, and approximate linear wavelet transforms. Due largely to these properties, Integer wavelet transforms are extremely useful for image compression systems requiring efficient handling of lossless coding, minimal memory usage, or low computational complexity, lossy-to-lossless recovery of images. Integer wavelet transform maps an integer data set into another integer data set. Integer wavelet transform (IWT) has the important property that IWT coefficients have the same dynamical range as the original signal. In discrete wavelet