# Genetic Algorithm

**Topics:**Steganography, Genetic algorithm, Optimization algorithms

**Pages:**11 (3320 words)

**Published:**February 11, 2013

A Secure Steganography Method based on Genetic Algorithm

Shen Wang, Bian Yang and Xiamu Niu

School of Computer Science and Technology Harbin Institute of Technology 150080, Harbin, China shen.wang@ict.hit.edu.cn; bian.yang@ict.hit.edu.cn; xiamu.niu@hit.edu.cn

Received April 2009; revised August 2009

Abstract. With the extensive application of steganography, it is challenged by steganalysis. The most notable steganalysis algorithm is the RS attack which detects the steg-message by the statistic analysis of pixel values. To ensure the security against the RS analysis, we presents a new steganography based on genetic algorithm in this paper. After embedding the secret message in LSB (least signiﬁcant bit) of the cover image, the pixel values of the steg-image are modiﬁed by the genetic algorithm to keep their statistic characters. Thus, the existence of the secret message is hard to be detected by the RS analysis. Meanwhile, better visual quality can be achieved by the proposed algorithm. The experimental results demonstrate the proposed algorithm’s eﬀectiveness in resistance to steganalysis with better visual quality. Keywords: steganography; steganalysis; genetic algorithm; RS algorithm

1. Introduction. Steganography is a branch of information hiding. It embeds the secret message in the cover media (e.g. image, audio, video, etc.) to hide the existence of the message. Steganography is often used in secrete communication. In recent years, many successful steganography methods have been proposed. Among all the methods, LSB (least signiﬁcant bit) replacing method is widely used due to its simplicity and large capacity. The majority of LSB steganography algorithms embed messages in spatial domain, such as BPCS[?, ?], PVD[?, ?]. Some others, such as Jsteg[?, ?], F5[?], Outguess[?, ?], embed messages in DCT frequency domain (i.e. JPEG images). In the LSB steganography, secret message is converted into binary string. Then the least signiﬁcant bit-plane is replaced by the binary string. The LSB embedding achieves good balance between the payload capacity and visual quality. However, the LSB replacing method ﬂips one half of the least-signiﬁcant bits. Thus the artifacts in the statistics of the image are easy to be detected[?]. Steganalysis is the method to reveal the hidden messages, even some doubtful media. The attacks on LSB replacing methods are most based on Chi-square analysis[?] and the relationship of pixels or bit–planes[?]. In the frequency domain, there are some steganalysis algorithms based on histogram and block eﬀect[?]. Among the methods, the RS steganalysis[?], proposed by Fridrich, is considered as the most reliable and accurate method to the LSB–replacing steganography. It utilizes the regular and singular groups as the statistics to measure the relationship of pixels. In most nature images, strong correlation exists in adjacent pixels. After the LSB–replacing steganography, the correlation is decreased. Thus, the proportion between the regular and singular groups changes and 28

A Secure Steganography Method based on Genetic Algorithm

29

the existence of the steganography is detected. Moreover, the secrete message length can be estimated by the amount of regular and singular groups. To resist to RS analysis, the inﬂuence on the correlation of pixels needs to be compensated. The compensation may be achieved by adjusting other bit planes. Nevertheless, the implementation may be computational infeasible. For example, if only two bit planes are modiﬁed in a 256 × 256 gray level image, there are 22 possible bit selections for each pixel. For the entire image, there are 2524288 times of adjustments. It is not feasible in the practical application. For this reason, optimization algorithms have been employed in information hiding to ﬁnd the optimal embedding...

Please join StudyMode to read the full document