Natias: Neuron Attribution based Transferable Image Adversarial Steganography

TL;DR

Natias introduces a neuron attribution-based method to enhance the transferability of adversarial steganography, effectively deceiving various steganalytic models and improving security against retraining defenses.

Contribution

The paper proposes a novel neuron attribution approach to corrupt critical features, significantly improving transferability of adversarial steganography across different models.

Findings

Enhanced transferability over previous methods.

Improved security in retraining scenarios.

Seamless integration with existing frameworks.

Abstract

Image steganography is a technique to conceal secret messages within digital images. Steganalysis, on the contrary, aims to detect the presence of secret messages within images. Recently, deep-learning-based steganalysis methods have achieved excellent detection performance. As a countermeasure, adversarial steganography has garnered considerable attention due to its ability to effectively deceive deep-learning-based steganalysis. However, steganalysts often employ unknown steganalytic models for detection. Therefore, the ability of adversarial steganography to deceive non-target steganalytic models, known as transferability, becomes especially important. Nevertheless, existing adversarial steganographic methods do not consider how to enhance transferability. To address this issue, we propose a novel adversarial steganographic scheme named Natias. Specifically, we first attribute the output of a steganalytic model to each neuron in the target middle layer to identify critical features. Next, we corrupt these critical features that may be adopted by diverse steganalytic models. Consequently, it can promote the transferability of adversarial steganography. Our proposed method can be seamlessly integrated with existing adversarial steganography frameworks. Thorough experimental analyses affirm that our proposed technique possesses improved transferability when contrasted with former approaches, and it attains heightened security in retraining scenarios.