Dual Watermarking Based on DCT with Human Visual Characteristics for Authentication and Copyright Protection

Nowadays, multimedia contents like images are easily distributed globally because of the widespread usage of information and communication technologies. Digital watermarking is an approach to prevent attacks or modification on images that might lead to serious problems like spreading of fake news, image piracy and illegal distribution of images. This research presents dual image watermarking based on human visual characteristics for authentication and copyright protection. The scheme aims to embed a watermark by modifying
$U$
coefficients of DCT-SVD with minimal distortion to the human eye. Additionally, the scheme demonstrates embedding authentication bits in the LSB for efficient tamper localization, enabling the detection of modifications made to the original image. The experimental results validate the effectiveness of the proposed scheme, showcasing superior performance in terms of imperceptibility, robustness, and tamper localization compared to existing watermarking schemes subjected to the same attacks. The scheme's ability to embed undetectable watermarks ensures copyright protection and authenticity verification, while its robustness against various image processing attacks. The proposed scheme offers a reliable approach for securing digital image content by ensuring imperceptibility, robustness, and tamper localization.