Sedenion polar harmonic Fourier moments and their application in multi-view color image watermarking

Abstract

In recent years, the research on image moments has made significant progress, transitioning from gray image moments to color image moments and from gray stereo image moments to color stereo image moments. However, the application of stereo image moments is limited to binocular stereo images with two views, and cannot handle images with more than two views. Multi-view images are more promising but are difficult to handle due to their intricate structure. Based on the sedenion theory and polar harmonic Fourier moments (PHFMs), this paper proposes the sedenion polar harmonic Fourier moments (SPHFMs) for processing multi-view color images. The imaginary part of sedenion encodes all views of multi-view color images, thus allowing all views to be processed as a single unit and internal correlations between all views to be maintained. The results of the stability experiment and image reconstruction experiment demonstrate that good stability and high image reconstruction performance are achieved when the SPHFMs are applied to processing multi-view color images. Since existing multi-view color image watermarking methods are not resistant to geometric attacks, we use SPHFMs for multi-view color image watermarking, which exhibits good imperceptibility and robustness.