A novel blind medical image watermarking scheme In NSST domain based on schur decomposition and DWT
Han Gao, Weimin Zheng · 2023
Nowadays, medical images face numerous information security challenges. On the one hand, medical images should be associated with patient information while ensuring data security and privacy protection. On the other hand, with the increasing use of technologies such as deep learning in medical images, there is a growing concern about copyright protection for medical image data used for sharing and training. Digital watermarking technology presents a potential solution to address these issues. This paper proposes a medical image watermarking scheme that utilizes Schur decomposition in combination with non-subsampled shearlet transform (NSST), discrete wavelet transform (DWT) and region of interest (ROI) identification for minimizing modifications to the original image. Furthermore, our proposed scheme does not require the original cover image for watermark extraction, reducing the reliance on additional information. The proposed method has been tested under various type of attacks, and the results have shown that it has high robustness and imperceptibility.