High Capacity Adaptive Block-Based Color Image Steganography and Human Perception-Guided Steganalysis

Abstract

In the realm of image steganography, where the delicate balance between secrecy and visibility is paramount, our proposed method represents a significant advancement. We leverage the human eye's heightened sensitivity to subtle changes in smoother image regions, introducing a novel approach to embedding secret data. Unlike conventional methods, our approach delves deep into the intricacies of each RGB plane in color images, optimizing data embedding while preserving visual integrity. Through a sophisticated block complexity measurement and selection process, we harness image redundancy efficiently using the Hilbert scan pattern. What distinguishes our method is its simplicity and self-sufficiency; no additional complexity is required for data extraction. Our experiments on benchmark datasets validate the superior stego image quality and precise data embedding capabilities of our approach. Moreover, our innovative human perception-based validation approach accounts for the complexities of the human visual system and subjective perceptions, enhancing the reliability of our method. This research not only advances the field of steganography but also lays the groundwork for more effective data hiding applications in color imagery.