It is important to restore original image and watermark in many applications such as military and medical. It is also referred to as ‘reversible watermarking’. Individual algorithm is difficult to analyze for a particular application, hence we compare the algorithm on some criteria. Here we present a comparative study of Difference Expansion (DE), Reversible Contrast Mapping (RCM) and Histogram Shifting (HS) methods. The results showed that HS method is the best method. It gave better PSNR than the other two methods, while embedding capacity of DE method proved to be better than the other two methods. We tested all methods for high and low quality images and found that all methods are suitable only for high quality images, while HS method is additionally suitable for cartoon and medical images. DE and RCM methods need improvement for applications in cartoon and medical images.

The techniques (Navnath and Rajendra, 2011) of the reversible watermarking can roughly be categorized into following types, namely, Difference Expansion (DE) (Tian, 2003; and Mohammad and Eran, 2011), histogram shifting (Tian, 2003; and Hsiang and Wan, 2011), contrast mapping (Dinu and Jean, 2007), integer wavelet transform, modulo 256 addition, lossless multi resolution transform, lossless compression, invertible noise adding, circular interpretation of bijective transformation, fuzzy and neural network. Other additional changes in reversible watermarking methods such as encryption (Liu and Liu, 2010) and coefficient adjustment (Chin et al., 2010) are also equally important to consider. Tian (2003) have presented watermarking using DE and embedding bit in least significant bit to achieve reversibility and very good embedding capacity. Alattar (Navnath and Rajendra, 2011) derived a DE transform for triplets and extended Tian’s algorithm to embed two bits in every triplet of pixels. Dinu and Jean (2007) developed a simple and efficient reversible hiding scheme based on Reversible Contrast Mapping (RCM), which does not need location maps to recover the host image. However, their scheme cannot fully control the position where information is embedded, so the quality of the watermarked image becomes poor. Histogram shifting method is one of the best methods of reversible watermarking. Wien et al. (2010) modified this method and introduced fast method.

Information Technology Journal, Reversible watermarking, Histogram Shifting (HS), Embedding capacity, Complexity, PSNR, Reversile Contrast Mapping (RCM), Difference Expansion (DE), Histogram Shifting, Reversible Contrast, Mapping, Expansion Methods .