Steganography: a class of secure and robust algorithms

TL;DR

This paper introduces a new class of steganographic algorithms that are secure and robust, leveraging chaos theory and formal security proofs, with detailed analysis in frequency domains.

Contribution

It presents a novel formalized approach to non-blind steganography using chaos theory, with proven security and robustness in frequency domains.

Findings

Algorithms are stego-secure against watermark-only attacks

Robustness is validated in DWT and DCT frequency domains

Formal security proofs support the approach

Abstract

This research work presents a new class of non-blind information hiding algorithms that are stego-secure and robust. They are based on some finite domains iterations having the Devaney's topological chaos property. Thanks to a complete formalization of the approach we prove security against watermark-only attacks of a large class of steganographic algorithms. Finally a complete study of robustness is given in frequency DWT and DCT domains.