An efficient lossless robust watermarking scheme by integrating redistributed invariant wavelet and fractional Fourier transforms

Abstract

An ensemble lossless watermarking scheme is proposed in the present study by integrating different concepts like redistributed invariant wavelet transform, discrete fractional Fourier transform, singular value decomposition (SVD) and visual cryptography within the framework of a single algorithm. The invariant wavelet transform helps to obtain the transform domain, which is invariant to flipping and rotation of image, this is followed by discrete fractional Fourier transform to obtain the translation invariant domain. Finally, embedding positions are selected based on a key and reliable features are extracted by performing SVD on a window centered at these positions. Based on these reliable features a binary map is generated through which a master share is created. The corresponding ownership share is produced from the master share and the watermark. In verification process the same operations of the embedding process are applied to the test image to obtain the master share and the watermark is recovered by stacking it over the ownership share. There are two main features of the proposed scheme (1) The quality of the image to be watermarked do not degrade during the process and (2) the extracted watermark can still be identified even from a seriously distorted image. These findings are also demonstrated with the help of a comparative study with several related schemes.