Dual Watermarking Using DCT-SVD for Image Authentication
TL;DR
The proposed dual image watermarking strategy aims to generate strong, accurate tampering localization while embedding both watermarked images and proof files into the host image and has achieved strong positive results against various attacks.
Abstract
With the innovation, development, and widespread use of today's communication technology, multimedia information such as images can be easily transmitted. Digital image watermarking is a technology that can prevent serious problems such as illegal image editing, fake news, and illegal image distribution. This research will introduce a new dual image watermarking strategy that uses human visual features to ensure image integrity, authorization, and protection. By using human perception features and adjusting the U coefficient in DCT-SVD, the proposed method further improves the existing watermarking process. It aims to generate strong, accurate tampering localization while embedding both watermarked images and proof files into the host image. According to the test results, the proposed scheme shows better performance in generating high uncertainty, robustness, and exposure identification ability compared to other methods when encountering similar images. In addition, the new dual image watermarking strategy provides a good solution for image authentication as it achieves an average sensitivity of 0.9837 and an accuracy of about 0.9836. The watermark quality of this solution reaches a PSNR of 48.41 and an SSIM of 0.9905. In addition, the proposed strategy has achieved strong positive results against various attacks. This solution provides a reliable way to protect digital images with the legal protection and integrity of the original images.