Reversible Watermarking Techniques: Digital Conent Security

Introduction

Now-a-days data tampering and copyright violation became a major challenge for digital content security. A watermarking technique is a promising solution for digital content protection. Apart from the watermarking there exists alternate solutions for digital content protection i.e., steganography and cryptography. Here, watermarking and steganography are considered to be data hiding techniques “it hides secret information into cover-content” (A.M. Alattar (2004), M. Barni et. al. (2004), B. Ou et. al. (2012), V MUNI SEKHAR, et. al. (2015). However, both have significant difference in terms of relationship between cover-content and secret information. Steganography is used for secret communication i.e., steganography conceals secret information into cover, if the existence of secret information is revealed, then steganography fails. In the case of watermarking the existence of secret information can be known to anyone, but separating it from cover-content is impossible. Whereas, cryptography does not conceal the secret information, but scrambles the information such that intruder cannot understand without an appropriate key [M. Barni et. al. (2004)].

A watermarking technique ensures protection of the data and ensures whether data is received with or without tempering. However, watermark can include additive noise in cover-content, so it may damage the sensitive information present in the cover. Hence, recovery of exact information at the receiving end may not be possible. In case of military, medical imaging and law-enforcement applications, even slight modification in cover-content may not allowed. Moreover, the need of such application is rapidly increasing every day. Hence, recovery of cover-content after extraction is an essential requirement for many applications. A lossless watermarking or reversible watermarking is a special case in watermarking [M.U. Celik et. al. (2005)]. It can extract exact original cover-content after the extraction process as shown in figure 1.

Figure 1.

Reversible watermarking process

The process of secret information embedding into cover-content is also called watermarking. From literature, based on sensitivity, watermarking techniques are classified into three categories such as semi-fragile, fragile and robust [H. He et. al. (2012), I. J. Cox et. al. (2008), V MUNI SEKHAR et. al. (2016)]. In semi-fragile watermarking technique, watermark is designed to survive minor modifications such as rotation and translation. Whereas, fragile not survive even minor modification. But, robust watermark can survive most image processing operations. Every type of watermark technique has its own set of applications. Here, most of the reversible watermarking techniques are fragile in nature. Even slight modification in watermarked data can be detected [M.U. Celik et. al. (2005), I. J. Cox et. al. (2008)].

For every watermarking technique have two important properties, one embedding capacity and other visual quality. The embedding capacity is denoting the maximum number of bits that can be embedded into cover-content. While, visual quality is the measure of similarity between cover-content and watermarking image measured with Peak Signal to Noise Ratio (PSNR). The performance of a reversible watermarking technique can be evaluated based PSNR and embedding capacity [[M.U. Celik et. al. (2005), I. J. Cox et. al. (2008), V MUNI SEKHAR et. al. (2013)].

Reversible watermark rapidly evolving from last two decades and this continuous evaluation not only gave more scope for designing robust authentication algorithms and also created a lot of confusion to the young scientists to take up reversible watermarking as a research problem. In this book chapter, we are intended to provide latest comprehensive review on reversible watermarking, it methodology and strengths. Moreover, the detailed review covers (through block-diagrams, tabular-comparisons, and actual performance-comparisons). Furthermore, different categories of reversible watermarking were reported in literature and it is hard to draw a boundary among different categories of watermarking. The common thing for all different approaches is finding space for data hiding.