A robust watermarking algorithm based on GLCM texture features, NSCT, and DWT
Haoxiang Yang, Zhongliang Pan · 2025
Balancing robustness, embedding capacity, and invisibility remains a key challenge in image watermarking. This paper proposes a robust watermarking algorithm combining Non-Subsampled Contourlet Transform (NSCT), Discrete Wavelet Transform (DWT), Singular Value Decomposition (SVD), and Gray-Level Co-occurrence Matrix (GLCM). NSCT is used to extract low-frequency coefficients from the host image, and the Arnold algorithm enhances watermark security. Host images are divided into 128×128 sub-blocks, with adaptive selection based on texture complexity analyzed via GLCM. Simulation results demonstrate the proposed algorithm's superior robustness against noise, cropping, filtering, and JPEG compression, while maintaining high embedding capacity and imperceptibility compared to existing methods.