Abstract

This study presents a robust and imperceptible blind watermarking system based on the frequency domain. Additionally, the approach is optimized using a meta-heuristic algorithm. The recommended technique involves applying the Discrete Wavelet Transform to the host picture to extract four sub-parts: diagonal, vertical, horizontal, and approximate. After that, the approximation sub-band is subjected to Singular Value Decomposition (SV-Decomposition). The optimal entrenching scaling factor used during the watermark (WM) implanting procedure is obtained by the SMO technique. Peak Signal-to-Noise Ratio (PSN-Ratio) assesses perceptibility quality, whereas Normalized Cross Correlation (NC) evaluates the robustness of the watermarked picture. Testing the presented process on watermarked images with a range of attacks yields decent PSN-Ratio and NC values.