Towards a tone mapping-robust watermarking algorithm for high dynamic range image based on spatial activity

Abstract

High dynamic range (HDR) imaging has received increasing attention due to its powerful capacity to represent real scenes as perceived by human eyes. However, studies on effective HDR image watermarking algorithms remain limited. In contrast to watermarking algorithms proposed for low dynamic range (LDR) images, several critical problems, such as the peculiar HDR floating-point number data format and various tone mapping operators (TMOs) that are widely used to adapt HDR images to conventional displays, need to be properly addressed. Hence, a novel HDR image watermarking algorithm robust against the effects of TMOs is proposed in this paper. Two important spatial activity concepts, i.e. the activity of robustness and activity of perception, are respectively defined to characterize the spatial diversity of the robustness and imperceptibility of tone-mapped images. Then, nonsubsampled contourlet transform and singular value decomposition are successively performed to extract the associated structural information which is invariable in HDR images and their corresponding tone-mapped images. In addition, hierarchical embedding intensity and a hybrid perceptual mask are designed to enhance the imperceptibility and robustness of the HDR image watermarking. Experiments with numerous HDR images and TMOs were conducted and the results show that the proposed algorithm is superior to the current state-of-the-art watermarking algorithms in terms of imperceptibility, robustness and embedding capacity.