Efficient Pairwise Reversible Data Hiding Technique Using in Image Authentication

I. INTRODUCTION Image Authentication is used to evidence that image is really what the user deems it is. For example, during image watermarking of patient's diagnostic image may occur errors in diagnosis and treatment, which may direction to possible life-blusterous outcome. Thus, to take over the problem of happening of artefacts and to making zero distorted or noise free watermarked medical images. Several reversible watermarking schemes have been initiated. Data hiding offering a way to situate data into cover medium for the motive of ownership fingerprinting, protection, secret communication authentication, and annotation [1]. In most data hiding algorithms, the cover data is ruinous abiding and cannot be exactly or same restored after the embedded message is extracted. At the latest, a new data hiding technique, that is, reversible data hiding (RDH) [2], is proposed. With the help of RDH the cover data and the embedded message both we can be extracted from the marked content. Reversible data hiding techniques wrap data in a host signal. For example, extraction allows of the original host signal and extracts an image also the embedded message. There are two main things are required for reversible data hiding techniques: the embedding capacity should be huge; and distortion should be minimum. These two requirements strife with each other. Generally, a higher embedding capacity results a higher degree of distortion. An improved Technique interpolates the same capacity with lower distortion or vice versa. Until now, various RDH algorithms have been proposed, e.g., compression based algorithms [3] [4], difference expansion (DE) based algorithms [5] [6] [7], histogram shifting (HS) based algorithms [8] [9], prediction-error expansion (PEE) based algorithms [10]–[15] and integer-to-integer transform based algorithms [16], some other RDH analysis have also been given etc. all of these techniques, PEE is widely applied nowadays because of its efficient capacity distortion trade-off. PEE was first proposed by Thodi and Rodriguez [10]. They were proposed that the difference between the pixel and its prediction is expanded for data embedding. When they compared between DE and HS based methods, PEE has a better performance. Hence the derived prediction-error histogram (PEH) is more rapidly distributed. Afterwards, Hu et al. [11] proposed PEE with an effective location map of smaller size, and thus growth of the capacity. Sachnev et al. [12] were proposed to sort prediction errors ethically to the local variance. Then prediction errors are process in the sorted order. Compared with former …