<title>Digital image watermarking on a special object: the human face</title>

Abstract

In this paper, we present a method for protection of digital contents by using the watermark embedding in special object, especially, human faces. To insert the watermark signals that are composed of noise like binary signals, we first localize the face regions within images by using the color and edge information. The skin color area is filtered out and then edge detector is applied for skin area to find out face features. These features are used for decision whether the skin area is face region or not. The face region is divide non-overlapping sub-blocks and a watermark bit is inserted into the each sub- block by considering the block activity. We insert a watermark bit in DCT domain of each sub-block. The level of modification of the DCT coefficients is determined considering the block variance. The non-zero coefficients of the DCT are selected and modified according to the robustness levels. Then, inverse DCT is performed. The extraction of the watermark is performed by comparing the original image in DCT domain. The robustness of the watermarking is similar to the other methods in DCT, but it has good visual qualities and less intended external piracy in terms of psychology.© (2000) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.