Block based visually secure image encryption algorithm using 2D-Compressive Sensing and nonlinearity

Abstract

In this paper, a block division based visually secure encryption algorithm is proposed for color images. Each block is sparsified through an optimal orthogonal transformation. Block based 2D-Compressive Sensing is applied to collect the compressed samples which best represent the signal, sampling is done through a measurement matrix (MM) which is generated by a 2D-SLIM map. MM is optimized through the proposed optimization algorithm which minimizes the mutual coherence between sparsified basis and MM. These compressed samples are further encrypted through a DNA XOR operation. The compressed-encrypted information is decomposed into the four sub images which are embedded into the host image through integer wavelet transform. SHA-256 is used to generate the seed value of the chaotic map. At the receiver end, the secret information is extracted with the correct knowledge of the keys. The numerical results and robustness of the proposed work is analyzed and the superiority of the developed algorithm is shown by comparing it with some recent published papers.