Triple scheme based on image steganography to improve imperceptibility and security

Abstract

A foremost priority in the information technology and communication era is achieving an effective and secure steganography scheme when considering information hiding. Commonly, the digital images are used as the cover for the steganography owing to their redundancy in the representation, making them hidden to the intruders. Nevertheless, any steganography system launched over the internet can be attacked upon recognizing the stego cover. Presently, the design and development of an effective image steganography system are facing several challenging issues including the low capacity, poor security, and imperceptibility. Towards overcoming the aforementioned issues, a new decomposition scheme was proposed for image steganography with a new approach known as a Triple Number Approach (TNA). In this study, three main stages were used to achieve objectives and overcome the issues of image steganography, beginning with image and text preparation, followed by embedding and culminating in extraction. Finally, the evaluation stage employed several evaluations in order to benchmark the results. Different contributions were presented with this study. The first contribution was a Triple Text Coding Method (TTCM), which was related to the preparation of secret messages prior to the embedding process. The second contribution was a Triple Embedding Method (TEM), which was related to the embedding process. The third contribution was related to security criteria which were based on a new partitioning of an image known as the Image Partitioning Method (IPM). The IPM proposed a random pixel selection, based on image partitioning into three phases with three iterations of the Hénon Map function. An enhanced Huffman coding algorithm was utilized to compress the secret message before TTCM process. A standard dataset from the Signal and Image Processing Institute (SIPI) containing color and grayscale images with 512 x 512 pixels were utilised in this study. Different parameters were used to test the performance of the proposed scheme based on security and imperceptibility (image quality). In image quality, four important measurements that were used are Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), Mean Square Error (MSE) and Histogram analysis. Whereas, two security measurements that were used are Human Visual System (HVS) and Chi-square (X2) attacks. In terms of PSNR and SSIM, the Lena grayscale image obtained results were 78.09 and 1 dB, respectively. Meanwhile, the HVS and X2 attacks obtained high results when compared to the existing scheme in the literature. Based on the findings, the proposed scheme give evidence to increase capacity, imperceptibility, and security to overcome existing issues.