Abstract: An image authentication scheme called watermarking is suggested to protect patients’ privacy by preventing patient documents from unauthorised access. In watermarking of medical image, in order to retain the quality of a region used for diagnostic purpose, known as Region of Interest (ROI), a prior knowledge about this area is needed. Most current medical watermarking schemes detect the ROI manually by assigned polygons or rectangles. In addition, in current automatic ROI detection algorithms, the robustness of ROI detection process is not being considered and addressed. It means the algorithms cannot tolerate any changes in the content of the medical images via any intentional or un-intentional attacks and the ROI detection algorithm is not efficient. By distributing and sharing the medical images between specialists and hospitals, they will be more vulnerable to different attacks. One of these potential attacks that may corrupt the medical images through transmission in communication channels is salt and pepper noise. In this paper an automatic ROI segmentation method for medical images watermarking application is proposed, which is robust against salt and pepper noise as a possible attack. Experimental results shows satisfactory performance of ROI segmentation under different densities of salt and pepper noise.

Abstract: Image Registration is a fundamental task in image processing used to match two or more pictures taken at different times and from different sensors. A various registration measurement have been developed for different types of problems. In this paper we present a novel registration approach based on normalized dissimilarity index which is resulting from local dissimilarity map that is a useful means to compare images. We compared its quality to classical registration measurements (Correlation and Mutual information) carried out on both binarized and grey-level images. We obtained good results for the two types of images. We showed afterwards the robustness of the proposed measure against the pepper and salt and Gaussian noise. We applied finally our registration approach on a medical images database on which we confirmed that the accuracy of alignment error estimation is improve compared to classical registration methods.

Abstract: In the digital world, due to the ease of creating, publishing and tampering of works, copyright protection became a major issue. This issue is much significant in movie industry which cause millions of dollars loses. With new technologies, almost everybody is able to produce, edit and share videos. Thanks to the Internet the sharing of works also, become a breeze. Hence, there should be a mechanism to prove the ownership and make sure of the integrity. Watermarking is one of the effective ways to help copyright protection. The aim of this work is exploring through the current watermarking solutions for copyright protection of digital video. For this purpose, foundations of video watermarking are explained firstly. Then a review on robust video watermarking methods for copyright protection application is proposed.