Predicted Pixel Pair Error Expansion based Reversible Embedding with Minimum Diamond Search Area

Prediction-error expansion (PEE) is a foundational contribution to the advancement of reversible data hiding (RDH). Prediction uses the image's first-order redundancy to conceal the secret data in a more condensed area. The precision of a predictor affects the impact of an RDH devise extensively. Nonetheless, the current PEE-based techniques, particularly several high-performing techniques rely solely on local knowledge for forecast or combination of errors in predictions. It is essential to expand the limited local area to leverage both local as well as non-local relationships synergistically. This study proposes a novel RDH approach utilizing diamond search. Based on the approach it was suggest exact slope path predictor , the initially estimated values are refined based to the correlations between the current pixel along with the referenced pixels, a process referred to as two-step prediction. The error after predicting the pixel might be associated directly towards the optimally matching pixel despite spatial constraints. Additionally, a novel complexity metric includes gradient as well as the deviation in average of contextual pixels is suggested. The integration of an proficient mapping technique with the multiple histogram modification method considerably enhances the visual appeal of the stego-image. Results demonstrate that the suggested two-stage prediction enhances payload, and the produced two-dimensional PEH exhibits a sharper distribution due to non-local pairing. Comprehensive experimental findings validated the efficacy of the proposed strategy in comparison to numerous leading-edge studies.