A Two Level Security Watermarking/Encryption Schema to Ensure the Protection of Medical Images
Sondes Ajili, Abdellatif Mtibaa · 2023
This research presents a novel watermarking technique specifically tailored for medical imaging applications. Our approach incorporates SVD, DCT, and DWT methods, complemented by AES to ensure the security and integrity of medical images while preserving their hidden information. The process begins with DWT applied to the medical image at the second level of decomposition. The resulting LL2 component is then divided into sub-bands, each sized $8 \times 8$. DCT is then applied to each sub-band, generating $8 \times 8$ blocks, which undergo SVD analysis. To achieve robust watermark extraction, we adopt an adaptive approach by customizing the watermark based on the SVD components. This adaptability ensures efficient retrieval of the watermark, even in the face of diverse types of attacks. To determine the appropriate “$\alpha$” for the watermark embedding process, we utilize the Weber constant. The proposed system undergoes rigorous testing against various attacks, including compression, noise, filtering, and geometric transformation. Through extensive experiments and analyses, we demonstrate the achievement of low distortion and high robustness in the watermarking process, while simultaneously ensuring enhanced security during the encryption phase.