A Study on Reversible Watermarking Using the Multi-level Scheme for Mosaic Images

This paper proposes a multiple layer combinations reversible watermarking scheme. We use a combination of predictors for calculating the new value of central pixel, as you can see in this paper we use a circular predictor, a cross one and also a square predictor, taking in consideration all the possibilities to mark the pixels. Also, an advantage of our method is the one of splitting the image into more than two sets of pixels, we create an advantage because we can cover better the image, and use precise predictors in different situations (we cover all type of situations – when we have shapes that must be recovered in decoding stage, but also when we have flat images). In the end, our results are obtained for mosaic images, we had this type of choice because using this type of test images we can do a division between the three channels of colors and can compute our algorithm much more in depth of details.