Linear weighted watermarking using normalized principal components

This manuscript introduces a novel linear weighted watermarking through normalized principal components using discrete wavelet transform (DWT) and singular value decomposition (SVD). Weight evaluation for embedding a singular value matrix of a watermark into a host is a trivial task in watermark embedding process. This task is accomplished by the normalized principal components derived from the singular value matrices of HH subbands of the host and the watermark images. Experiments are conducted to analyze the effectiveness of the proposed watermark embedding and watermark-host extraction processes against the other DWT–SVD based watermarking schemes. Performance of this method is analyzed by peak signal to noise ratio, structural similarity index and correlation coefficient. The proposed method is also tested against various geometrical and non-geometrical attacks on watermarked images.