Software model of a highly undetectable stegosystem – ⊕HUGO model

Abstract

The article is devoted to the research of a software model of the hidden message transmission process using one of the new methods of steganography proposed by the authors. The presented model implements a hard-to-detect method for embedding hidden messages in digital still images used as covering objects (containers). The main principle of using the model is to minimize adequately specified distortion using a new efficient encoding algorithm. Distortion is defined as an operation ⊕ (excluding OR - XOR) between the next half-byte of the hidden message and the right half-byte of the specially selected byte of the covering object, which is selected from the condition for minimizing the penalty function calculated on some selected neighbourhood of this byte. This allows authors to keep the original image almost unchanged and make it indistinguishable from the original, even after large payloads are embedded in it. Thus, the embedded hidden message may not be visible when the steganalysis even for the experienced steganalytic. The ⊕ operation has the property of bijectivity (it is reversible), which makes quite easy to use it to extract a hidden message on the recipient’s side, if the recipient has key information. The stegosystem model is called the ⊕Highly Undetectable SteGOsystem model. In short, it can be referred to as the ⊕HUGO model.