Hidden data embedding in uninformative subsets of subband image projections

The paper proposes a method of subband hidden data embedding into images, which can be applied to solve the urgent problem of control over the distribution and use of images in information and telecommunications systems. The developed method is based on changing the values of container image subband projections to the eigenvectors of cosine transform subband matrices corresponding to a given two-dimensional interval of spatial frequencies. In this paper, authors propose a decision rule for choosing uninformative subsets of subband projections that can be used for hidden embedding. Comparative computational experiments were carried out to estimate the distortions of container im-ages during data embedding based on the developed method and well-known methods of hidden em-bedding, such as the method of relative replacement of discrete cosine transform coefficients and the method of spectrum expansion. It is shown that the developed method of subband hidden embedding in some cases has an advantage in terms of the secrecy of data embedding into images and allows to embed data without causing significant distortions of container images. The method and decisive rule proposed in the article can be used in the development of an intelligent information system for the hidden data embedding into images, which allows achieving a high degree of data embedding secrecy based on making the decisions about the rational applied embedding parameters values.