A High-capacity HE Interpolation-based Reversible Data Hiding Method using Difference Interval-based Substitution

Abstract

With the growing emphasis on Internet data transmission security, an increasing number of scholars are exploring the field of data hiding. In the spatial domain, various techniques exist, such as pixel expansion, histogram shifting, and the application of interpolation techniques (IT). IT has garnered attention as a novel design approach, particularly in IT-based RDH (reversible data hiding), known for its ability to maintain relatively high image quality while enhancing hiding capacity compared to other methods.
However, previous research has predominantly focused on crafting interpolation images that closely mirror the original image quality, followed by proposing error prediction-based data hiding strategies. Departing from this conventional path, this study introduces the concept of single reference point diffusion to propose the homogeneous expansion-based interpolation (HEI) algorithm and then applies the HEI algorithm to generate interpolation images less prone to offset pixel effects. Subsequently, the difference interval substitution (DIS) strategy is employed to maximize hiding capacity while preserving favorable image quality. Experimental results validate that our approach significantly boosts hiding capacity, nearly tripling that of the Zhang et al.’s method under comparable computational complexity, while maintaining an average image quality of 31.54 dB.