Multiple scaling Fuzzy-PSO watermarking scheme for gray-scale and colored images

Digital watermarking has become the core of digital security paradigm for ownership rights and copyright protection. The advance research in digital watermarking domain is focused on developments of techniques and algorithms which produce optimum results balancing two most relevant watermarking parameters, visual quality and robustness. This paper proposes a novel watermarking scheme which is a hybrid of fuzzy logic with particle swarm optimization (PSO), and is applied over eight standard grayscale and colored images each, to analyze the effectiveness of the proposed watermarking scheme. Three human visual system (HVS) features i.e. luminance, edge sensitivity, and contrast sensitivity, are extracted from each host image in the discrete cosine transform (DCT) domain, which are used as fuzzy attributes. These features are fed as input into fuzzy inference system which is based on 27 inference rules. Nine different image processing attacks are applied over signed images to examine the robustness of the proposed algorithm. Multiple scaling factors have been used for the embedding strength (∝) in the proposed scheme. The quality of the signed and attacked images is assessed using computed peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) values. The similarity between the extracted and the original watermark is assessed by computing normalized cross-correlation (NC) values depicting the robustness of our proposed scheme. The obtained results are compared with other state-of-the-art schemes available in this domain. The proposed Fuzzy-PSO watermarking scheme is found to outperform all of them.