Multi-Contextual Design of Convolutional Neural Network for Steganalysis

TL;DR

This paper introduces a novel multi-contextual CNN architecture with learned denoising and self-attention for improved steganalysis, effectively capturing diverse embedding zones and enhancing detection accuracy.

Contribution

It proposes a new multi-contextual CNN with learned denoising and self-attention modules, addressing the challenge of detecting diverse embedding zones in steganalysis.

Findings

Outperforms prior steganalysis methods in accuracy.

Learned denoising kernels improve noise residual quality.

Self-attention enhances focus on embedding-prone regions.

Abstract

In recent times, deep learning-based steganalysis classifiers became popular due to their state-of-the-art performance. Most deep steganalysis classifiers usually extract noise residuals using high-pass filters as preprocessing steps and feed them to their deep model for classification. It is observed that recent steganographic embedding does not always restrict their embedding in the high-frequency zone; instead, they distribute it as per embedding policy. Therefore, besides noise residual, learning the embedding zone is another challenging task. In this work, unlike the conventional approaches, the proposed model first extracts the noise residual using learned denoising kernels to boost the signal-to-noise ratio. After preprocessing, the sparse noise residuals are fed to a novel Multi-Contextual Convolutional Neural Network (M-CNET) that uses heterogeneous context size to learn the sparse and low-amplitude representation of noise residuals. The model performance is further improved by incorporating the Self-Attention module to focus on the areas prone to steganalytic embedding. A set of comprehensive experiments is performed to show the proposed scheme's efficacy over the prior arts. Besides, an ablation study is given to justify the contribution of various modules of the proposed architecture.