Robust hybrid watermarking for security of medical images in computer-aided diagnosis based telemedicine applications
Ramsha Ahmed, Bilal Hassan, Bo Hu Li · 2018
Telemedicine has gained much popularity in recent times. It refers to the use of state of the art telecommunication and information technologies to provide medical and health care services to distant locations. However, secure transmission and access of the medical information is a critical issue. Medical data theft or altering the medical information of a patient is a rising and dangerous crime. Therefore, the authenticity and verification of a patient's medical information is vital as they form the basis of examination. This paper proposes a robust medical image watermarking technique to provide authenticity and verification of medical images in Computer-Aided Diagnosis (CAD) based telemedicine applications. To achieve high robustness and imperceptibility with watermark transparency, a hybrid technique based on Fast Curvelet Tansform (FCT) and Robust Principal Component Analysis (RPCA) is structured. Curvelet transform gives optimal sparse representation in comparison to most traditional transforms and aids in efficient watermark recovery. At receiver end, the extracted watermark provides authentication and verifies the medical image under examination. The simulation results show that the proposed technique provides good imperceptibility as the average value of Peak Signal-to-Noise Ratio (PSNR) is around 51.9537 dB and average value of Structure Similarity Index (SSIM) metric is around 0.9981. In addition, the proposed technique is more robust against various image processing techniques and common attacks such as noise addition, filter, rotation etc. in terms of Normalized Correlation Coefficient (NCC) and Bit Error Rate (BER).