A New Data Hiding Scheme Combining Genetic Algorithm and Artificial Neural Network

A New Data Hiding Scheme Combining Genetic Algorithm and Artificial Neural Network

Ayan Chatterjee (Sarboday Public Academy, India) and Nikhilesh Barik (Kazi Nazrul University, India)
DOI: 10.4018/978-1-5225-2857-9.ch005
OnDemand PDF Download:
No Current Special Offers


Today, in the time of internet based communication, steganography is an important approach. In this approach, secret information is embedded in a cover medium with minimum distortion of it. Here, a video steganography scheme is developed in frequency domain category. Frequency domain is more effective than spatial domain due to variation data insertion domain. To change actual domain of entropy pixels of the video frames, uniform crossover of Genetic Algorithm (GA) is used. Then for data insertion in video frames, single layer perceptron of Artificial Neural Network is used. This particular concept of information security is attractive due to its high security during wireless communication. The effectiveness of the proposed technique is analyzed with the parameters PSNR (Peak Signal to Noise Ratio), IF and Payload (bpb).
Chapter Preview


Steganography is the art of hiding secret data or secret information at the time of wireless communication (2013). In this special approach, the existence of communication among sender and intended receiver(s) can be hidden from unintentional receiver(s) or hacker(s). In the particular system of methodology, the secret information is embedded in a cover medium, such as- image, audio, video etc. and the embedded file is transferred through communication channel. In Image steganography, image pixels are used for inserting secret data (2014). In video steganography, video frames are used to embed the data. Each video frame is treated as an image. The fact is that video files are safer than image files, because- video files take a large no. of image pixels over image files. So, obviously, insertion of data in video files consists of less distortion over image files. There are two major categories of steganography- spatial domain and frequency domain. In spatial domain steganography, the secret data is inserted directly in the image domain.

In frequency domain steganography, the actual domain of cover medium is converted to another medium using some mathematical transformation. This transformed domain is used for inserting secret data. Comparing these two different categories, it is observed that the distortion between actual cover medium and stego medium is generally less in spatial domain steganography over frequency domain steganography. In other words, peak signal to noise ratio (PSNR) and MSE (Mean Squared Error) generally give better result in spatial domain. But it can be easily hacked by unintended receiver(s) using pseudo random number generator (PRNG). Frequency domain steganography is better than spatial domain with respect to PRNG and other statistical attacks. Different mathematical transformations, such as- Discrete Cosine Transformation (DCT), Discrete Wavelet Transformation (DWT), Discrete Fourier Transformation (DFT), Fast Fourier Transformation (FFT) etc are used to transform the actual domain of cover medium. Also, in data compression, different stochastic optimization schemes are used. Among them, Genetic Algorithm (GA), Fuzzy logic etc are very important. Genetic Algorithm (GA) is basically a soft computing based optimization approach. But, the hereditary properties of animal, the concepts of which are used in this particular approach, are very much effective in various fields rather than optimization. Among them, image processing, information security, artificial intelligence etc. are very much important. Basically, GA is developed with three different properties- selection of chromosomes, crossover and mutation. According to the variation of these three operations, various types of GA are developed. GA is generally very much effective for NP hard problems. Another important tool in soft computing is Artificial Neural Network (ANN). In engineering field, this is a sequence of patterns like neurons of human being. In generally, at the time of implementation of all soft computing based approaches, given information and corresponding ingredients are taken as input. Target or goal is obtained Output in all the approaches. The speciality of the approach ANN is that the source pattern and target are taken as input and corresponding ingredients are obtained as output. This speciality makes ANN more effective than other schemes. Depending on variation of source pattern and target, different ANN schemes are developed. Among them, single layer perceptron, multi layer perceptron etc. are very much important. In this paper, a particular approach of frequency domain steganography is developed by using crossover operator of GA and single layer perceptron of ANN. In the next section, some related works are discussed. Then the proposed data hiding scheme is illustrated followed by the algorithm. After that the efficiency of the scheme is analyzed with different type experiments. At last, conclusion of the scheme is given followed by future direction of the work.

Key Terms in this Chapter

Peak Signal to Noise Ratio (PSNR): An expression is used to realize the distortion of quality of a cover medium at the presence of noise during wireless communication.

Steganalysis: Security analysis system at the time of sending data through wireless communication by the schemes of steganography.

Video Steganography: Process of authenticated communication by hiding secret information from unauthorized user(s) through a video file as cover medium.

Wireless Communication: Communication process with independent of wire among a finite set of users, who are in a long distance.

Artificial Neural Network (ANN): An important tool to set up linkage between provided input and required output and it is developed on the basis of the set up of communicating nervous system of human being.

Genetic Algorithm: A special algorithmic optimization procedure, developed on the basis of simple hereditary property of animals and used for both of constrained and unconstrained problem. In Artificial Intelligence (AI), it is used as heuristic search also.

Complete Chapter List

Search this Book: