Abstract:Recently, due to the considerable feature expression ability of neural networks, deep linguistic steganalysis methods have been greatly developed. However, there are still two issues that need to be ameliorated. First, the prevailing linguistic steganalysis methods rely heavily on massive training data, which is labor-intensive and time-consuming. Second, these methods implement steganalysis only in different weak-concealed scenarios, the stego texts in each of which have only a single language style and payload. But in practice, the intercepted network samples are probably the mixture of the stego texts that possess different language styles and payloads, in which the semantic spatial distribution may be more chaotic than that in weak-concealed scenarios, thus making steganalysis more difficult. To address the above issues, a novel linguistic steganalysis method is proposed in this letter. First, the pre-trained BERT language model is constructed as an embedder to compensate for the shortage of data. Then, in addition to learning local and global semantic features, a feature interaction module is designed for exploring mutual effects between them. Furthermore, besides the typical cross-entropy loss, triplet loss is also introduced for the model training. In this way, the proposed method can refine more comprehensive and discriminative deep features in the intricate semantic space. The performance of the proposed method is compared with the representative linguistic steganalysis methods on datasets of different scales, and the experimental results reveal the superiority of the proposed method.

SeSy: Linguistic Steganalysis Framework Integrating Semantic and Syntactic Features

A New Steganalysis Approach Based on Both Complexity Estimate and Statistical Filter

Research on Holism-Based Feature Extraction and Fusion for Steganalysis

Universal Methodology for Developing Quantitative Steganalysis

Hybrid Dictionary Learning for JPEG Steganalysis

Linguistic Steganography: from Symbolic Space to Semantic Space

Linguistic Steganalysis Toward Social Network

Linguistic Steganalysis with Graph Neural Networks

Small-Scale Linguistic Steganalysis for Multi-Concealed Scenarios

Linguistic Steganalysis Via Densely Connected LSTM with Feature Pyramid

High-Performance Linguistic Steganalysis, Capacity Estimation and Steganographic Positioning.

Enhancing Steganographic Text Extraction: Evaluating the Impact of NLP Models on Accuracy and Semantic Coherence

State-of-the-art Advances of Deep-learning Linguistic Steganalysis Research

LINK: Linguistic Steganalysis Framework with External Knowledge

Hi-Stega: A Hierarchical Linguistic Steganography Framework Combining Retrieval and Generation

A Fast and Efficient Text Steganalysis Method

Generative adversarial networks-based image steganography with multiscale features integration

TS-CNN: Text Steganalysis from Semantic Space Based on Convolutional Neural Network

Linguistic Steganalysis in Few-Shot Scenario

Towards Next-Generation Steganalysis: LLMs Unleash the Power of Detecting Steganography

Semantic-Preserving Linguistic Steganography by Pivot Translation and Semantic-Aware Bins Coding