Position based watermarking

Paulo Vinicius, Paulo Borges, J. Mayer · 2004

This paper proposes a block based watermarking scheme where the information is associated to watermark blocks positions in a host image. Considering a digital image of size N/spl times/M there are many possibilities (given by the combination expression) as how to combine a given number of watermarks. For each of these possibilities, it can be assigned a bit string, making use of a coding scheme. The detector searches in the image for k positions that present the highest correlation to watermark blocks, determining where in the image these blocks were embedded. The decoding process relates these positions to a bit string. Also evaluate this technique regarding perceptual impact and show that for the same number of inserted bits use less energy if compared with other watermarking methods.

Read the paper · More papers on PaperTik