A novel robust image watermarking in quaternion wavelet domain based on superpixel segmentation (Q2211698)

Quaternions Wavelet Transform (QWT) is a new kind of multi-resolution analysis tool of digital image processing, and it is near shift-invariant. The QWT is an extension of the real wavelet transform and the complex wavelet transform (CWT) based on quaternion algebra and the 2-D Hilbert transform of filter theory. The CWT overcomes the real wavelet transform and CWT limitations when used for a watermarking scheme. Using superpixel image segmentation and QWT, the authors propose an approach for digital image watermarking. The proposed algorithm is highly robust against common image processing operations and local geometric transformations. Experimental results are provided to illustrate the efficiency of the proposed image watermarking, especially for noise attacks and local desynchronization attacks. The present work is new, interesting, and can be chosen for practical use.