File-Size Preserving LZ Encoding for Reversible Data Embedding

Abstract

Methods for recycling the redundancy due to reference multiplicity in the LZ77 algorithm have already been proposed. These methods can be characterized in terms of reversible data embedding. In these methods, the redundancy in LZ77 is used to embed extra information in codewords. This paper proposes an LZ77 variation that specializes in reversible data embedding. The proposed encoding algorithm performs neither compression nor expansion. Instead, it embeds maximum possible extra information without changing the input file size. The asymptotic embedding capacity of this algorithm is evaluated, and it is shown that a duality relation exists between compressibility and embedding capacity.