Transform Domain and Singular Value Decomposition for Robust and Imperceptible Image Watermarking
Muhammad Fikri Aufa, Ledya Novamizanti, Rita Purnamasari · 2022
As technology advances at an ever-increasing rate, data is disseminated online through multimedia, particularly images, which are relatively simple to access. However, this situation is prone to theft of digital data and copyright infraction. A method to preserve and maintain multimedia data information, particularly images, is the robust watermarking approach. The implanted watermark can be recovered even after being altered by numerous techniques and will be difficult to erase. The host image is altered by redundant wavelets, and the suggested watermarking approach incorporates watermarks situated in the low and high-frequency sub-bands of the curvelet transform to provide a fair trade-off between robustness and imperceptibility. Applying a discrete cosine transform to the watermark causes it to be integrated into the image. Then, using the singular value decomposition approach, the watermark’s unique value is added to the host image’s unique value. The watermarked image is extracted on the receiving side in order to return the watermark to its original state. According to the study’s findings, the scheme has a good imperceptibility and is resilient with a maximum peak signal to noise ratio value of 71.7630 dB, SSIM 1, NC 1, and BER 0. The suggested approach is resistant to attacks from Gaussian filter, rotation, and JPEG compression with noise that is (Gaussian, salt-and-pepper, speckle).