Introduction
In recent years, a special kind of digital watermark, called reversible (which is also called lossless, invertible, erasable, distortion-free, etc.) data hiding technique, has drawn lots of interest. It not only declares the ownership of the digital media by embedding the digital watermark into the cover image but also can completely restore the original image after extraction. This kind of data hiding is suitable for some specific applications where images are sensitive to further processing, such as medical image, satellite image, and artwork.
In 2001, Honsinger proposed the first lossless data hiding concept providing lossless authentication. However, image quality made by the scheme was unsatisfactory. It suffered from salt-and-pepper visual artifacts. Then Vleeschouwer used patchwork algorithm and circular histogram to solve the noise problem. Although this scheme could survive high quality JPEG compression, its embedding capacity was very limited. Three years later, Ni proposed a reversible data hiding method based on histogram modification. The secret data was embedded by shifting the histogram. It demonstrated that the lower bound of the PSNR of a watermarked image is 48.13 dB. Additionally, lossless compression is also used for reversible data hiding. Fridrich compressed the least significant bit plane of selected pixels using lossless compression, and then combined the compressed result with the watermark to produce the bitstream. Then the bitstream was embedded into the cover work by using LSB substitution. Their algorithm achieved reversibility, but its payload was very limited. In 2002, Fridrich proposed another cheme to improve the embedded payload and visual quality. Each disjointed group of the cover ork was categorized into one of three sets: regular (R), singular (S), and unusable (U). The group types R and S were used to embed 1 and 0, respectively. During the embedding procedure, a lipping function was used to flip the embedded group into the specific set. Finally, the sets R and S were losslessly compressed and embedded into the cover work. In 2005, Celik proposed an improved scheme by using generalized-LSB embedding. The scheme first converts binary to L-ary, then lossless compression (CALIC) is used to compress the quantization residues as side information. This scheme efficiently uses CALIC to obtain high embedding capacity.
For some applications such as satellite and medical images, reversible data hiding is the best solution to provide copyright protection or authentication. Being reversible, the decoder can extract the hidden data and recover the original image without distortion. In this project, a reversible data hiding scheme based on prediction error expansion is proposed. The predictive value is computed by using various predictors. The secret data is embedded in the cover image by exploiting the expansion of the difference between a pixel and its predictive value. Experimental results show that our method is capable of providing a great embedding capacity without making noticeable distortion. In addition, the proposed scheme is also applicable to various predictors.
In recent years, a special kind of digital watermark, called reversible (which is also called lossless, invertible, erasable, distortion-free, etc.) data hiding technique, has drawn lots of interest. It not only declares the ownership of the digital media by embedding the digital watermark into the cover image but also can completely restore the original image after extraction. This kind of data hiding is suitable for some specific applications where images are sensitive to further processing, such as medical image, satellite image, and artwork.
In 2001, Honsinger proposed the first lossless data hiding concept providing lossless authentication. However, image quality made by the scheme was unsatisfactory. It suffered from salt-and-pepper visual artifacts. Then Vleeschouwer used patchwork algorithm and circular histogram to solve the noise problem. Although this scheme could survive high quality JPEG compression, its embedding capacity was very limited. Three years later, Ni proposed a reversible data hiding method based on histogram modification. The secret data was embedded by shifting the histogram. It demonstrated that the lower bound of the PSNR of a watermarked image is 48.13 dB. Additionally, lossless compression is also used for reversible data hiding. Fridrich compressed the least significant bit plane of selected pixels using lossless compression, and then combined the compressed result with the watermark to produce the bitstream. Then the bitstream was embedded into the cover work by using LSB substitution. Their algorithm achieved reversibility, but its payload was very limited. In 2002, Fridrich proposed another cheme to improve the embedded payload and visual quality. Each disjointed group of the cover ork was categorized into one of three sets: regular (R), singular (S), and unusable (U). The group types R and S were used to embed 1 and 0, respectively. During the embedding procedure, a lipping function was used to flip the embedded group into the specific set. Finally, the sets R and S were losslessly compressed and embedded into the cover work. In 2005, Celik proposed an improved scheme by using generalized-LSB embedding. The scheme first converts binary to L-ary, then lossless compression (CALIC) is used to compress the quantization residues as side information. This scheme efficiently uses CALIC to obtain high embedding capacity.
For some applications such as satellite and medical images, reversible data hiding is the best solution to provide copyright protection or authentication. Being reversible, the decoder can extract the hidden data and recover the original image without distortion. In this project, a reversible data hiding scheme based on prediction error expansion is proposed. The predictive value is computed by using various predictors. The secret data is embedded in the cover image by exploiting the expansion of the difference between a pixel and its predictive value. Experimental results show that our method is capable of providing a great embedding capacity without making noticeable distortion. In addition, the proposed scheme is also applicable to various predictors.
No comments:
Post a Comment