A novel reversible data hiding in encrypted technique for medical image using image interpolation and a two-stage difference preservation encryption
TL;DR
This work proposes a novel VRAE framework based on interpolation-based difference, creating a stable embedding space via a specialized encryption design that achieves a significantly higher embedding capacity on multiple test images compared to existing state-of-the-art schemes.
Abstract
With the rapid development of information technology and network transmission, the generation, storage, and sharing of images are widely applied in daily life. Reversible Data Hiding in Encrypted Images (RDHEI) is a key technology addressing security and privacy challenges. However, existing Vacating Room After Encryption (VRAE) methods for achieving high capacity often rely on MSB prediction, a technique whose performance is inherently tied to image texture. To overcome this, we propose a novel VRAE framework based on interpolation-based difference, creating a stable embedding space via a specialized encryption design. Our methodology uses a unique two-stage encryption process, applying Difference-Preserving Encryption (DPE) to downsampled reference pixels, followed by applying a pseudorandom constant mask (PRF Mask) to non-reference pixels. The core innovation is that the same constant mask is applied to both the original and interpolated images, precisely preserving the difference structure. This enables a "capacityadaptive" embedding strategy that yields bits from zero-difference blocks to secret data, maximizing efficiency. Experimental results confirm that the proposed method achieves a significantly higher embedding capacity on multiple test images compared to existing state-of-the-art schemes. Security analysis confirms the method's robustness against statistical attacks, while complete reversibility is guaranteed.