FFT-ETM based distortion less and high payload image steganography

Abstract

This paper is presenting a novel high capacity based imperceptible and robust image steganography technique for obscured communication. A considerable literature studied on this domain reveals distortion that drastically affects image quality. These techniques obscure covert data in most significant bits or least significant bits of host image via easy or unsystematic replacement. Such schemes are vulnerable to malevolent attacks like sample pair method, chi-square test, and quality of host image especially badly affected by MSB replacement. Furthermore, such schemes are lacking in carrying maximum covert information as the number of host image pixels and covert image pixels has the ratio 8:1. In our proposed scheme robust and imperceptibility feature is injected using insignificant pixel value divergence of host and a high capacity covert image. We have proposed frequency entropy method that compares frequencies of covert image and host image in FFT (Fast Fourier Transform) domain. The eminent rate of frequency ETM (Entropy Threshold Match) leads to good image quality and information carrying capability. Moreover, our proposed technique also encrypts the secret image in frequency domain with multi flipped permutated random key vector that provides robustness. Therefore, experiments exhibit that this scheme has improved signal to noise ratio and BPP (bits per pixel) in contrast to existing schemes.