Parametric Binary Tree Labeling Based RDHEI Technique

Abstract

The current paper proposes a reversible data hiding technique in an encrypted image (RDHEI) using a parametric binary tree labeling (PBTL) scheme. The PBTL scheme provides a range of prediction errors to be utilized in data embedding. The binary tree is characterized by a tuple parameter, which is configured to provide the highest embedding rate. Yi et. al.’s technique introduces PBTL based RDHEI scheme, but it lacks high embedding capacity and high security. Further, it utilizes a causal predictor, which is not efficient in predicting embeddable errors. The proposed technique suggests a two-layer embedding strategy in encrypted image in which one of the pixel layers is exploited using the non-causal predictor and another pixel layer is exploited using the causal predictor. The advantages of both causal predictor and non-causal predictor are utilized in the proposed technique to increase embedding capacity than the conventional techniques. In addition to this, data security is also improved by employing a standard pixel-based encryption technique. Thus, in comparison to conventional techniques, the proposed technique maximizes the embedding rate and also ensures high data security.