Advantages And Disadvantages Of Watermarking

Recently due to internet growth, lots of medical digital images are being shared between medical experts and hospitals for better and more precise diagnosis,as well as for research and educational purpose and also for many other commercial and non-commercial applications. In modern times now, all the integrated health care systems like Hospital Information System (HIS) and Picture Archiving and Communication System (PACS) allow easy distribution of medical imageswherethe whole or the region of interest(ROI) of the image can be manipulated or debased by hackers without great effort.Besides, the huge clinical data like demographic data, images and reports whichare transmitted betweenhealth care systems should be stored, processed and controlledwith
plus radiologist or specialist’s inference could be embedded into medical images as watermark data so that error probability decreases. Watermarking methods could be classified into blind and non-blindbased on the information required at the detector to extract the watermark. In blind methods, which are also known as public watermarking schemes and are the most challenging type of watermarking systems, neither the host image nor watermark content are needed to extractthe watermark whereas in non-blind methods, which are also called private watermarking schemes, both the host image and embedded watermark contents are required to extract the watermark. It is clear that the non-blind methods offer better robustness compared to blind methods, but in the case when it is hard to get the original cover and watermark image forwatermark extraction, the blind methods are more desired. The watermarking schemes can alsobe categorized as spatial domain or frequency domain techniques depending on embedding of watermark contents into host image is done either in spatial or in transform
In this scheme,DWT is applied on ROI of the host image to get different frequency sub-bands of its wavelet decomposition. SVD is applied on the non-overlapping blocks of the LLsub-band with size 4 × 4 and the watermark contents are embedded into the elements in second and third rows offirst column of left singular matrix U since they have much closer value and represent proper threshold. The values of this pairare reclaimed using the threshold to embed a bit of watermark content.One binary image (logo) and other text data have been used as watermark contents.Thisscheme offers better visibility of watermarked image and recovery of watermark content due toDWT-SVD combination.Using proposedscheme, watermark contents are extracted successfully under various noise attacks like JPEGcompression, filtering, Gaussian noise, Salt and pepper noise, cropping, filtering and rotation.Performance comparison of proposed scheme with existing schemes shows proposed schemehas better robustness against different types of