Robust Message Embedding via Attention Flow-Based Steganography

TL;DR

A novel message embedding framework, called Robust Message Steganography (RMSteg), which is competent to hide message via QR Code in a host image based on an normalizing flow-based model and is the first work that integrates the advantages of transformer models into normalizing flow.

Abstract

Image steganography can hide information in a host image and obtain a stego image that is perceptually indistinguishable from the original one. This technique has tremendous potential in scenarios like copyright protection, information retrospection, etc. Some previous studies have proposed to enhance the robustness of the methods against image disturbances to increase their applicability. However, they generally cannot achieve a satisfying balance between the steganography quality and robustness. Instead of image-in-image steganography, we focus on the issue of message-in-image embedding that is robust to various real-world image distortions. This task aims to embed information into a natural image and the decoding result is required to be completely accurate, which increases the difficulty of data concealing and revealing. Inspired by the recent developments in transformer-based vision models, we discover that the tokenized representation of image is naturally suitable for steganography task. In this paper, we propose a novel message embedding framework, called Robust Message Steganography (RMSteg), which is competent to hide message via QR Code in a host image based on an normalizing flow-based model. The stego image derived by our method has imperceptible changes and the encoded message can be accurately restored even if the image is printed out and photoed. To our best knowledge, this is the first work that integrates the advantages of transformer models into normalizing flow. Our experiment result shows that RMSteg has great potential in robust and high-quality message embedding.

Figures (27)

• Figure 1: Compared with previous methods that can only embed limited bit-level information, RMSteg can achieve a much higher embedding capacity and meanwhile has better steganography quality. Also, it can survive various real-world distortions.
• Figure 2: The pipeline of RMSteg. We first transform the QR Code encoded with the secret message to make it easier to hide through an invertible neural network (a). After that, we perform invertible token fusion (ITF) (b) on the tokenized QR Code. We then use a normalizing flow-based model with attention affine coupling blocks (AACBs) to implement data concealing and revealing (c). During training, we employ a distortion simulation module (d) to simulate real-world image disturbances.
• Figure 3: Some QR Code transition results, the transformed QR Codes are still identifiable.
• Figure 4: Stego images and decoded QR Codes under different distortions. QR Codes with green borders can be recognized while those with red borders cannot. Zoom in for better observation.
• Figure 5: Stego images generated by different methods.