Design and implementation of a block-based lossless watermarking scheme

Abstract

High capacity reversible watermarking method based on block difference expansion is presented. The proposed scheme relies upon redundancy analysis for image authentication and tamper localisation. The performance of such scheme is compared with that of Tian. While Tian's algorithm is a pixel-based embedding scheme, the proposed method is a block-based embedding technique. Thus, it can be used with the following advantages: (1) possibility of providing more capacity for authentication data (2) no compression is needed (as in case of Tian's scheme) and (3) low time cost because of its reasonable complexity. To find out the embedding capacity both information and redundancy are examined for the underlying image. In addition a corresponding difference image is defined, its histogram is determined and its entropy and redundancy are computed. The proposed scheme creates an extra space for the authentication payload by making use of block difference expansion.