Article Preview
TopIntroduction
Image hiding is a kind of information hiding known as steganography that is usually used in convert communication. The difference between steganography and cryptography is that the goal of cryptography is to make the secret data unreadable by a third party (attackers); while steganography aims to hide secret data within cover data. Depending on the embedding domains, image steganographic methods are classified into two categories: spatial domain based methods and frequency domain based methods. The most popular steganography methods that fall into the spatial domain class are presented in (Adelson, 1990; Bender, Gruhl, Morimoto & Lu, 1996; Bender et al., 2000; Turner, 1989). The basic idea of these methods is to use the Least Significant Bits (LSBs) of image pixels to embed the secret data. For example, if the value of secret data is 5 (101)2, and the pixel value of the cover image is 100 (01100100)2, then the basic method is to embed the secret data into the pixel value in the three LSBs to be 109 (01101101)2. The human eyes cannot detect such changes since the Most Significant Bits (MSBs) have not changed. Despite that the spatial domain based LSB methods are simple and have less computational complexity, they have weak resistance to attacks (Cheddad, Condell, Curran & Mc Kevit, 2010). In order to overcome this shortcoming and improve the security level, frequency domain based methods have been proposed. The main idea of this class of image steganographic methods is to convert the image into frequency domain using a transformer and embed the secret data there. There are several transformation methods that have been used for steganography. The most popular methods are the Discrete Cosine Transformation (DCT) method that was used in (Chang, Chen & Chung, 2002; Hashad, Madani & Wahdan, 2005), Discrete Fourier Transform (DFT) method that was used in (Mckeon, 2007; Raja, Chowdary, Venugopal & Patnaik, 2005), and Discrete Wavelet Transform (DWT) method that was used in (Xuan et al., 2002a; Xuan et al., 2002b). Further image hiding methods are discussed and analyzed in (Cheddad et al., 2010).
The efficiency of image hiding process can be improved by augmenting it with other approaches such as searching algorithms. One of the current approaches that has recently been adapted in order to improve image hiding is Genetic Algorithm (GA), which was proposed by Holland in 1962 (Davis, 1991; Goldberg, 1989). It is a search technique that is used to generate optimal or approximate solutions for a specific problem. The GA uses terms stimulated by biology’s terms, such as inheritance, mutation, selection, and crossover. Basically, the algorithm starts from a population of randomly generated individuals, and the fitness for each individual in the population is evaluated. Then, based on the resultant fitness, new individuals will be mixed using one or more crossovers, randomly mutated, and finally selected to be a new generation. The new population will be used in the next iteration of the algorithm that will terminate by either reaching a maximum number of generations or by reaching a specific fitness level, depending on a defined objective function.
On the other hand, Optimal Pixel Adjustment Process (OPAP) (Chan & Cheng, 2004) is a technique that is used to reduce the distortion that might be caused by the LSB method. The OPAP is carried out after embedding the secret image where each pixel of the stego-image (containing the secret image) is adjusted to reduce the difference between the stego-image pixel value and the corresponding cover image pixel value in order to improve the quality of the stego-image.