Enhancing Image Watermarking: An Innovative Multi-Objective Genetic Algorithm-Based DWT-SVD Approach for Robustness and Imperceptibility

This paper investigates the enhancement of the image watermarking algorithm through its robustness and imperceptibility.We propose a watermarking method for protecting image data that is established using an optimal Discrete Wavelet Transform and Singular Value Decomposition (DWT-SVD).A multi-objective genetic algorithm (MOGA) with two conflicting objectives (i.e., PSNR and Hamming Distance (HD)) is employed to minimize the embedding distortion and maximize the robustness.These goals are achieved by merging natural selection and GA, producing a powerful tool for coefficients embedding optimization.GA has been used to guide the selection process of DWT coefficients for watermark embedding.We use Arnold transform to scramble the watermark bits to increase the watermark security.Consequently, various evaluation functions such as Peak signal-to-noise and SSIM are calculated to examine the watermarked image quality.Eventually, the selected coefficients represent the optimal choices to minimize the embedding distortion and maximize the robustness against attack.The final results of experiments demonstrate that the presented method is robust to many types of attacks, namely, Salt & Pepper, Gaussian, Speckle, Poisson, Resizing, Rotation, Median filter, cropping, and compression.The findings demonstrate that the suggested method performed better by implying embedding distortion and attack resistance than the current techniques.