A Technique for Self-embedding of Digital Images

One of the promising ways to provide robustness of digital images transmission is self-embedding technique. A motivation of self-embedding is to provide robustness of the image against some sorts of tampering, to provide an ability to detect tampering and to recover the original image if possible. The first subject of this paper is a new format of image lossy compression and its use for represen-tation of the image version to be self-embedded. The approach is based on using singular value de-composition (SVD) of the image matrix, which makes it possible to represent an image as a partial sum of the most significant layers (MSLs). Based on this idea, a new format for coding of the image to be embedded is developed. The second subject of the paper is self-embedding task and recovery algorithm of the possibly tampered image. A peculiarity of the developed algorithm of recovery is that it makes possible to detect the corrupted blocks and a possibility to recover them automatically. The software tool implementing the proposed approach is developed and made use for validation of the proposed self-embedding technique