Multiple watermark embedding scheme in wavelet-spatial domains based on ROI of medical images

Watermarking in medical images is a new area of research. It has the potential of being a value-added tool for medical confidentiality protection, patient-related information hiding, and information retrieval. Medical image watermarking requires extreme care when embedding additional data within the medical images because the additional information must not affect the image quality as this may cause misdiagnosis. In this paper we present a scheme that depends on the extraction of the ROI (region of interest) and its use as a watermark to be embedded twice; first as a robust watermark in the RONI (region of non interest) in the wavelet domain and again as a fragile watermark in the ROI in the spatial domain. Moreover multiple watermarks such as the physician's digital signature and EPR (Electronics Patient Record) are embedded in the RONI in wavelet domain depending on a private key. We compare this scheme by another one that we presented before to show the robustness of new scheme. In our work we use MRI brain images with a brain tumor as the ROI. The experimental results showed that the watermarked image is robust to JPEG compression, ROI removal, and addition of an additional tumor to the image and some geometrical attacks; lowpass and median filtering and some types of noise as; Gaussian, Poisson, Salt and Pepper and finally Speckle.