Abstract:

Signal processing transformations and mathematical techniques are generally used in watermarking techniques used in areas such as copyright protection, authentication, fingerprinting, content tagging. In this study, instead of the Singular Value Decomposition (SVD), which is preferred in most watermarking schemes, the dimension reduction-based truncated-SVD technique is used. This technique is combined with the Discrete Wavelet Transform. Compared to the baseline SVD-DWT-based technique, it has been observed that the proposed scheme has made progress in imperceptibility and resistance performances against all possible attacks, except histogram equalization. It is predicted that the proposed scheme will lead to alternative stamping schemes using different matrix decomposition and signal transformations.

Keywords: