WSOMedMark: Robust and optimized dual image watermarking using RDWT and its authentication using BRISK and MSER features for smart healthcare system

Information contained in digital medical images is beneficial for patient’s diagnosis. Medical Image Watermarking (MIW) provides precise patient identification and reduces the possibility of errors during diagnosis. A combination of Redundant Discrete Wavelet Transform (RDWT) and Discrete Cosine Transform (DCT) is applied to host medical images. QR decomposition and Randomized Singular Value Decomposition (RSVD) techniques are used to embed a logo and text image. Chaotic logistic mapping is used to encrypt the watermarks before embedding in order to enhance security. The optimum value of the strength factor is obtained by using White Shark Optimization (WSO). The performance of the proposed scheme is evaluated in terms of robustness, embedding capacity, imperceptibility and computational complexity. Performance comparison with existing schemes shows the superiority of the proposed scheme. The average percentage improvement of Peak Signal to Noise Ratio (PSNR) is 29.34% and of Structural Similarity Index Measure (SSIM) is 2.11% compared to related schemes. Authentication is performed by utilizing Binary Robust Invariant Scalable Keypoints (BRISK) and Maximally Stable Extremal Regions (MSER) feature detectors. Important features of watermarked images are preserved even after applying several attacks and both the feature detectors successfully extract these features from attacked watermarked images.