Computation of embedding capacity in reversible watermarking schemes

Abstract

This paper deals about embedding capacity computation for reversible watermarking schemes. The paper proposes a unique way of computing embedding capacity directly from the data set without actually embedding the watermark in the image. This computation is done based on the statistical parameters of the data set. We also demonstrate how to compute the capacity under distortion constraints. We also show how to enhance the capacity by using a multipass embedding scheme without substantially affecting the PSNR.

Pixel-QUAD based Reversible Data Hiding for High Embedding Capacity

“…For obtaining higher embedding capacity, researchers devised multi-pass embedding technique, wherein watermarking is done in consecutive passes to the cover image and then to the watermarked image. The author has addressed this subject in [3]. All this techniques are based on Pixel-Pair approach.…”

“…A theorem is proposed in this regard in [3]. Thus the over error after n passes of watermarking, would be (3∆) n .…”

Estimation of the Embedding Capacity in Pixel-pair based Watermarking Schemes

“…the pixel pair (10,12), s [1] as (9,14), s [2] as (5,18) and so on. Similarly, i s[0] , i s [1] and i s [2] are 0,0 and φ respectively and η(i s[0] ), η(i s [1] ) and η(i s [2] ) are 1,1 and 0 respectively.…”

“…These techniques are location map based and hence require an additional step of data compression. Recently reversible contrast mapping (RCM) based methods [2,5,6] have been used to efficiently embed data without using a location map. Prediction based techniques [7,19] have also been suggested and they use information from the neighboring pixels to embed information.…”

“…This is a class of transforms which operate on independent groups or blocks of pixels, generally known as the expansion and mapping based algorithms. In this paper, we propose algorithms for pixel-pair based methods, since for majority of these techniques [2,5,7,6,11,14,22], the independent group or blocks of pixels are pixel pairs. In other words the entire image is partitioned into disjoint pairs of pixels.…”