Steganalyzer performances in operational contexts

TL;DR

This paper investigates the performance of steganalyzers in real-world scenarios, examining their robustness to parameter changes and the potential of combining multiple methods to improve detection accuracy.

Contribution

It explores the feasibility of creating universal steganalyzers without prior knowledge and evaluates the impact of parameter variations and method merging on detection performance.

Findings

Performance degrades with parameter modifications between training and testing.

Combining multiple methods during learning can enhance detection accuracy.

Universal steganalyzers without specific knowledge are challenging but possible.

Abstract

Steganography and steganalysis are two important branches of the information hiding field of research. Steganography methods consist in hiding information in such a way that the secret message is undetectable for the uninitiated. Steganalyzis encompasses all the techniques that attempt to detect the presence of such hidden information. This latter is usually designed by making classifiers able to separate innocent images from steganographied ones according to their differences on well-selected features. We wonder, in this article whether it is possible to construct a kind of universal steganalyzer without any knowledge regarding the steganographier side. The effects on the classification score of a modification of either parameters or methods between the learning and testing stages are then evaluated, while the possibility to improve the separation score by merging many methods during learning stage is deeper investigated.