Abstract
The paper presents an improved digital image watermarking algorithm by incorporating the discrete Fourier transform (DFT) and singular value decomposition (SVD). The Fourier transformed carrier image is decomposed into four different frequency subbands by the proposed onion peel decomposition (OPD) algorithm and the SVD-based watermarking scheme is applied to attach the transformed watermark in all four carrier subbands. The proposed inverse OPD algorithm together with inverse DFT is utilized to reconstruct the watermarked image from the frequency blocks. The watermark extraction algorithm is simple and it performs the inverse of watermarking process. The experimental analysis on different images shows that the proposed algorithm produces good quality watermarked images. The watermarks extracted from watermarked images inflicted with potential attacks are also of better perception quality than those produced by other prominent algorithms in terms of subjective and objective metrics.