A Genetic Algorithm based steganography in frequency domain (GASFD)

Abstract

Image steganography is the art of hiding information onto the cover image. In this paper a Genetic Algorithm based color image authentication/data hiding technique through steganographic approach, in frequency domain using Discrete Fourier Transform (DFT) termed as GASFD, has been proposed. 2×2 masks are taken from the source image in row major order where DFT is used to transform original image (cover image) block from spatial domain to frequency domain. Three bits of the hidden image are embedded per byte of the source image onto the rightmost 3 bits of each pixel excluding the first byte of each mask, as a result large volume of message/ image is embedded in frequency domain. 2×2 embedded image mask is transformed from frequency domain to spatial domain using inverse DFT. Resulting image mask of size 32 bits are taken as initial population. New Generation and Crossover are applied on the initial population to obtain stego image. In the process of embedding dimension of the hidden image followed by the content of the message/hidden image are embedded. Reverse process is followed during decoding. Genetic algorithm is used to enhance the security level. Various statistical parameters computed are compared with the existing Discrete Cosine Transformation based steganographic algorithms devised by Hashad A. I. et al. [9] which shows that proposed GASFD obtained better results in terms of large message embedding and consistent PSNR.