Journal of Data Science and Modeling

A Hybrid Image Watermarking Method Based on Hessian Matrix Eigenvalues and Wavelet Transform for Digital Copyright Protection

Articles in Press

Document Type : original

Authors

Hadis Heidari
mosayyeb amini

Kermanshah University of Technology

10.22054/jdsm.2026.94577.1103
Abstract
The widespread exchange of images over the internet and the growing threats of unauthorized copying and content manipulation have made digital image watermarking essential for protecting intellectual property and ensuring image authenticity. A major challenge in designing watermarking algorithms is achieving imperceptibility while preserving image quality. This paper proposes a hybrid digital image watermarking approach based on Hessian matrix eigenvalues and the discrete wavelet transform. After decomposing an image into frequency sub-bands using the wavelet transform, regions with stable structural and textural information are identified based on the Hessian matrix eigenvalues, and suitable coefficients for watermark embedding are selected. Combining spatial features derived from the Hessian matrix eigenvalues with the multi-scale analysis capability of the wavelet transform enables effective watermark distribution within the image. The proposed method is evaluated using peak signal-to-noise ratio, structural similarity index measure, and normalized correlation. Experimental results demonstrate that the method preserves watermark imperceptibility, making it suitable for security applications, digital copyright protection, and image authenticity verification.

Keywords

Digital image watermarking
Discrete wavelet transform (DWT)
Hessian matrix eigenvalues
Copyright protection
Subjects
Journal of Data Science and Modeling

Articles in Press, Accepted Manuscript
Available Online from 09 November 2026

Home

Guide for Authors

Submit Manuscript

Contact Us