Singular value decomposition (SVD)-based watermarking scheme is a well-known copyright protection technique due to its orthogonal vectors that represent geometrical information about an image and its singular values (SVs) with exceptional stability. Unfortunately, this scheme is succumbed to false-positive problem (FPP), where an adversary is able to steal the copyright of the digital media and dispute ownership with the content’s owner. Researchers discovered that the \(U\) and \(V\) orthogonal vectors of SVD, which store the geometrical information of an image, are the main cause of FPP. By using either \(U\) , \(V\) or both orthogonal vectors in the embedding process, FPP can be avoided; however, this approach is not ideal because \(U\) and \(V\) vectors are hypersensitive to even a little change in their matrices, negatively affecting the imperceptibility of the watermarked image. Furthermore, most watermarking algorithms chose the watermark embedding strength using single scaling factor (SSF). SSF, however, is not an ideal value that can strike the right balance between robustness and imperceptibility. Therefore, in this paper, existing watermarking schemes that make use of orthogonal vectors of SVD were analyzed, and then a new hybrid SVD-based image watermarking scheme using human visual system (HVS), discrete wavelet transform (DWT) was proposed. Particle swarm optimization (PSO) algorithm is used with the proposed scheme to find the best scaling factor that strikes a compromise between robustness and imperceptibility. The experimental results showed that the proposed scheme performs better in terms of robustness, to most types of geometrical assaults and image processing than existing schemes, while still producing good-quality watermarked images. These findings highlight the effectiveness of the proposed approach in strengthening digital watermarking techniques, making it a promising solution for secure and reliable copyright protection in various digital applications.
