Secure ‘Text in Image’ Steganography Scheme Based on Alexnet and Contour-Let Transform

Abstract

Nowadays, ‘text in image’ steganography is utilised in a variety of applications like military, surveillance, and remote sensing etc., in order to keep the secret information secure. This paper is presenting the unique steganography algorithm for improving data security, visual quality, quality metrics and withstand the attacks of the stego image. In proposed algorithm, the security of the confidential information is improved by utilizing the SLICE encryption algorithm. It is used to generate the cipher data from the confidential data. The alexnet is introduced in this paper to detect the facial area in human cover image. A total of 1987 images are used to train the network and got maximum accuracy. Contour-let transform is utilized to decomposition of alexnet output. The “random pixel embedding” (RPE) technique is utilized to embed the confidential data in facial area of the sub-band. The combination of alexnet and contourlet transform is used to generate good visual quality of the stego image. The proposed algorithm produces a stego image that has better visual quality, security, metric-values and withstand attacks compared to recent methods.