Encryption‐Based Twining of Visible and Invisible Watermarking for Secure Transmission of Medical Images
Rajat Sood, Anamika Shukla Sharma, Ashima Anand · Security and Privacy · 2025
ABSTRACT The modern healthcare system demands the safeguarding and dependability of data due to the widespread usage of digital medical records and the exchange of confidential patient information. Preserving the confidentiality and integrity of medical images is of utmost importance as they play a vital role in diagnosing, devising treatment plans, and conducting research. Nevertheless, the inherent digital aspect of these images renders them readily accessible, modifiable, or susceptible to misuse. To mitigate these dangers, the healthcare industry is placing growing significance on the use of robust watermarking systems to safeguard patient data. This article proposes a robust dual watermarking framework that integrates conspicuous and imperceptible watermarks using a combination of Non‐Subsampled Shearlet Transform (NSST), Bi‐dimensional Empirical Mode Decomposition (BMEMD), and Multi‐scale Singular Value Decomposition (MSVD). The approach embeds a visible ownership watermark and a hidden authentication watermark while enhancing security through Redundant Discrete Wavelet Transform (RDWT)‐Randomized Singular Value Decomposition (RSVD)‐based encryption. Experimental evaluation on a subset of 3000 images from the NIH Chest x‐ray dataset demonstrates strong robustness against various attacks such as Gaussian noise, JPEG compression, and blurring. The proposed method achieves a 47.49% improvement in Normalized Correlation (NC) over existing techniques and maintains high visual quality, as indicated by PSNR of 71.4742 dB and SSIM of 0.9997. These results highlight the method's potential for secure medical image transmission and authentication compared to existing watermarking schemes.