Optimal DWT-SVD Domain Image Watermarking Using Multi-objective Evolutionary Algorithms

In this paper, we present an optimal Discrete Wavelet Transform-Singular Value Decomposition (DWT-SVD) based image watermarking scheme using Pareto-based Multi-Objective Evolutionary Algorithm (MOEA). After decomposing the cover image into four bands, we apply the SVD to each band, and the singular values of each sub-band of the host image are modified by different scaling factors to embed the watermark image. Modifications are optimized using a fast Elitist Nondominated Sorting Genetic Algorithm (NSGA-II) to obtain the highest possible robustness without losing the transparency. Since the same watermark with proper scaling factors is embedded in 4 blocks, it is extremely difficult to remove or destroy the watermark from all frequencies. Experimental results show improvement both in transparency and robustness under a wide range of attacks.