Securing the regions comprising information in a document image with novel key selection technique – a parallel approach

Abstract

Securing the information contents in an image with increase in speed is a challenging issue even today. In this paper, a parallel approach to secure information contents of important regions in a document image is proposed. The novelty of the approach are to identify intelligently the regions comprising information is to be secured, generating and selecting the key sequence. The proposed method comprises of four stages. In the first stage, rules of cellular automata are selected randomly to generate the key sequence. In the second stage, key sequences are selected on the basis of sarpabandha which depicts pattern of the movement of snake. The input color image is decomposed into non overlapping blocks of uniform size which are processed in parallel. Using Grey Level Co-occurrence Matrix (GLCM), entropy feature is extracted for each block and based on the feature value classification of blocks are carried out in the third stage. In the last stage encryption operation is performed for each plane of input color image separately. Red, green and blue color planes of input color image are fused with the key sequences to perform the encryption. Cipher is feedback to enhance the diffusion of information. The proposed methodology is analyzed in terms of parameters peak signal to noise ratio, entropy, structural similarity index and correlation coefficient. In the proposed work, identification of regions and encryption both are processed in parallel to reduce the time of computation.