Steganalysis of Content-Adaptive Steganography in Spatial Domain

Abstract

Abstract. Content-adaptive steganography constrains its embedding changes to those parts of covers that are difficult to model, such as textured or noisy regions. When combined with advanced coding techniques, adaptive steganographic methods can embed rather large payloads with low statistical detectability at least when measured using feature-based steganalyzers trained on a given cover source. The recently proposed steganographic algorithm HUGO is an example of this approach. The goal of this paper is to subject this newly proposed algorithm to analysis, identify features capable of detecting payload embedded using such schemes and obtain a better picture regarding the benefit of adaptive steganography with public selection channels. This work describes the technical details of our attack on HUGO as part of the BOSS challenge.

Steganalysis of LSB Replacement Using Parity-Aware Features

“…For such embedding operations, the most accurate detectors today are built as classifiers using features obtained as sampled joint distributions (co-occurrence matrices) among neighboring elements of noise residuals [12,11,27,25,13]. These detectors perform equally well for both LSB replacement and LSB matching because features formed from noise residuals are generally blind to pixels' parity.…”

Breaking HUGO – The Process Discovery

“…In this section, we only provide a rather brief description, referring to [14] and our other paper in this volume [9] for a more detailed exposition of this methodology, experimental evaluation and comparison to SVMs as well as a discussion on the relationship of our approach to prior art in machine learning.…”

“…This rule of thumb does NOT hold when the cover-source mismatch is absent. Without the mismatch, the detection accuracy simply keeps on improving with increased feature dimensionality (see our other paper [9] in this volume). The rest of our record submissions are displayed in Fig.…”

“…As a result, we learned quite a bit and we certainly hope that the reader of this paper will as well. This paper accompanies another paper [9] in this volume, which contains additional technical details of our final approach together with an extensive experimental section.…”

“…This information is usually not found in technical papers that only show the final polished approach. This work accompanies our other paper in this volume [9]. …”