An adaptive pixel value ordering based reversible data hiding scheme for images

Pixel value ordering (PVO) is a prominent method for reversible data hiding, which has gained increasing research interest in recent years due to its effective performance producing high-fidelity marked images. PVObased methods performance lies on block dimensions and thresholds for classification. However, most methods determine proper parameters by performing exhaustive searches in restricted search spaces. In this paper, we present a novel Pixel Value Ordering technique based on Genetic Algorithms (PVO-GA) to address these issues. PVO-GA uses flexible sized blocks to take advantage of different block features, to better exploit image redundancy. A genetic algorithm is introduced to optimize the embedding process since the solutions space is greatly expanded. Results showed that PVO-GA achieves satisfactory levels in terms of the introduced distortion while significantly reduces the computational cost needed to find the optimal set of insertion parameters when compared to previous methods.