Compressive sensing on transform domain-based medical images watermarking

Preservation of patient data is one of the crucial issues in the field of health care. An electronic patient record (EPR) is a computer-based database of health data. The EPR must be protected from unauthorized parties, errors in patients diagnosing and verifying the authenticity of data. Watermarking is a method of inserting information data such as EPR into a digital data. In this study, we propose Compressive Sensing (CS) method to be applied to transform domain-based medical image watermarking. Those method are applied to medical image watermarking based on two transform domains, namely Fast Discrete Curvelet Transform (FDCuT) and Discrete Cosine Transform. CS can secure and also reduce the size of the watermark. The technique of inserting a watermark into the host uses the Singular Value Decomposition (SVD) method by replacing the diagonal matrix on the host with a watermark diagonal matrix. This research has a good performance of imperceptibility and robustness with PSNR values of 57.3295 dB, SSIM 0.9998, NC 1, and BER 0. This watermarking scheme is also resistant to JPEG compression attacks, noise addition, signal processing, and geometry attacks. Therefore, this watermarking scheme can protect patient medical records safely and has a high level of robustness.