An imperceptible semi-blind image watermarking scheme in DWT-SVD domain using a zigzag embedding technique

Abstract

This research work presents a semi-blind image watermarking scheme based on transforms domain DWT-SVD using an efficient embedding technique. This scheme provides a high level of robustness and imperceptibility for digital image copyright protection. In the embedding process, the watermark data (logo and image in this research) is first transformed into the frequency domain using the DWT algorithm for one pass. Then, the LL-band values are also transformed by the SVD algorithm. S values (diagonal matrix) are prepared for inserting into the cover images. Differently, the cover images (SD and HD quality) are also transformed by DWT for two levels of decomposition. During the insertion process, the watermark bits are embedding into the HL2 and HH2 bands of the cover image using the zigzag technique to improve the imperceptibility. Moreover, the obtained watermarked images are subjected to different attacks (geometric, image processing, and jpeg compression) to show the robustness of the proposed scheme. Finally, the extraction process needs only the watermarked image and the U and V values of SVD transforms (semi-blind watermarking) to reconstruct the original watermark. The performance parameters for the robustness and imperceptibility involved in this work are PSNR and Normalized Correlation Coefficient NCC metrics. Computational cost is also calculated for both embedding and extracting watermark process for different cover image types. Test results indicate that the proposed scheme has a better performance when the watermark data is embedding in a zigzag way.