Robust Steganography for Online Secure Communication with Binary Text Image in JPEG Compressed Domain

Abstract

This paper introduces a novel steganography method of secure communication to mitigate the perceptual degradation associated with the quantization process in JPEG compression, particularly when images are recompressed at standard quality levels by potential attackers. Our approach operates within the compressed domain, optimizing the selection of cover images based on the presence of high-texture blocks, thereby enhancing robustness and capacity while avoiding visual artifacts. This technique ensures a high Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM) without the common compromise on visual quality. Additionally, our method allows for the retrieval of embedded messages without the need for the original image, making it highly applicable to real-time communication scenarios. Through extensive experimentation, we demonstrate that cover images containing over 80% textured blocks, with blocks selected for embedding having at least two non-zero quantized Discrete Cosine Transform (DCT) coefficients beyond the DC component, significantly improve PSNR values over existing methods while maintaining high payload capacity. The system exhibits robustness against JPEG recompression across a wide range of quality factors (42 to 99) and resilience to various image processing attacks, marking a significant advancement in the field of image compression and secure communication.