Abstract
In the field of Reversible Data Hiding in Encrypted Images (RDHEI), significant efforts have been dedicated to improving embedding efficiency, privacy, and accurate reconstruction. However, the synchronization between the embedded data and the cover media, which is crucial for ensuring robustness, has received relatively limited attention. This paper introduces an innovative RDHEI method based on the range asymmetric numeral system (rANS) to enhance both robustness and embedding performance. For this, the proposed rANS-RDHEI method brings forward multiple advancements. First, a strategic block transformation technique is implemented to reorganize the image blocks, positioning high-frequency symbols prominently within each block to optimize the rANS encoding process. Additionally, a novel progressive frequency aggregation approach is applied, allowing iterative block compression without storing full symbol frequencies, thereby minimizing storage overhead and freeing substantial embedding space. Additionally, an optimal encoding endpoint selection mechanism, guided by consecutive loss analysis, determines the encoding cutoff per block, enhancing the rANS encoding efficiency. This enables the proposed RDHEI process to reserve substantial space within the image, which is then used for embedding through an efficient bit-replacement technique. Experimental evaluations reveal that the rANS-RDHEI method outperforms existing and related RDHEI techniques in terms of embedding capacity while achieving ∞ PSNR. Moreover, the proposed method has shown high resistance against statistical and perceptual attacks with high synchronization capability.