Watermarking in halftone image with maximal child node dependence coefficients in wavelet domain

To achieve high image quality, this paper proposes a novel approach but efficient algorithm for halftoning watermarking. Wavelet coefficients of edge details have strong relations, while the dependence of coefficients of texture is more weak. On the other hand, large wavelet coefficients contain the information of image brink. Considering the relationship of intra-scale and maximal coefficients, the boundary error measure is built. While the gray image turns to halftone image using proposed algorithm, binary watermark is embedded in the edge position. In order to enhance the ability to resist cropping and smearing attack, the watermark image is pre-treated with Arnold transformation processing. The proposed method is effective, because the halftone image is embedded as a watermark sequence during the halftone process. Experiment results show that this algorithm does not cause significant distortion of the image and is good resistance to cropping, daub, JPEG compression and noise attacks.