Generally, the embedding quality and the embedding capacity are two important
factors in evaluating the performance of information hiding schemes, but they
are usually inversely proportional. Thus, to satisfy the different requirements, the users
often make a tradeoff between good embedding quality and high embedding capacity. In
this paper, a spatial domain technique, embedding binary message bits into a still image
based on Hamming code oriented wet paper codes, is discussed and evaluated. We get
inspiration from Chang, Chou and Lin’s work and make a critical improvement. The significant
difference between their scheme and ours is that we have no specific requirements
imposed on the distribution of both the dry pixels and the wet pixels in cover images; wet
pixels can be chosen completely in accordance with the characteristics of different cover
images as well as the application’s needs. The proposed scheme has been tested with nine
commonly-used gray images. Subtle and radical variations to the existing scheme are
suggested and proven to provide a significantly improved PSNR.
Relation:
International Journal of Innovative Computing, Information and Control