Effective Message Hiding with Order-Preserving Mechanisms

2024-02-29

British Machine Vision Conference

AI Summary

To address the low recovery accuracy caused by message bit-order loss and the trade-off between capacity and imperceptibility in image steganography, this paper proposes StegaFormer. It introduces the first Order-Preserving Message Encoder (OPME) and Decoder (OPMD) to explicitly model bit-sequence structure. A Global Message-Image Fusion (GMIF) mechanism is designed, leveraging cross-attention for deep cross-modal semantic alignment. The framework adopts a pure MLP architecture with end-to-end differentiable encoding and decoding. Evaluated on COCO and DIV2K, StegaFormer achieves over 8.2% higher bit recovery accuracy and 23% greater payload capacity than state-of-the-art methods, while attaining SSIM ≥ 0.992—demonstrating superior imperceptibility. The source code is publicly available.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Message hiding, a technique that conceals secret message bits within a cover image, aims to achieve an optimal balance among message capacity, recovery accuracy, and imperceptibility. While convolutional neural networks have notably improved message capacity and imperceptibility, achieving high recovery accuracy remains challenging. This challenge arises because convolutional operations struggle to preserve the sequential order of message bits and effectively address the discrepancy between these two modalities. To address this, we propose StegaFormer, an innovative MLP-based framework designed to preserve bit order and enable global fusion between modalities. Specifically, StegaFormer incorporates three crucial components: Order-Preserving Message Encoder (OPME), Decoder (OPMD) and Global Message-Image Fusion (GMIF). OPME and OPMD aim to preserve the order of message bits by segmenting the entire sequence into equal-length segments and incorporating sequential information during encoding and decoding. Meanwhile, GMIF employs a cross-modality fusion mechanism to effectively fuse the features from the two uncorrelated modalities. Experimental results on the COCO and DIV2K datasets demonstrate that StegaFormer surpasses existing state-of-the-art methods in terms of recovery accuracy, message capacity, and imperceptibility. We will make our code publicly available.

Problem

• Achieving high recovery accuracy in message hiding
• Preserving sequential order of message bits
• Addressing modality discrepancy between messages and images

Innovation

• MLP-based framework preserves bit order
• Segments message into equal-length sequential segments
• Cross-modality fusion mechanism integrates uncorrelated features