Image Multi-frequency Watermarking based on Human Visual System

C.-T. Hsieh, Y.-K. Wu, Y.-L. Lu, and C.-P. Luo (Taiwan)


watermark, DWT, HVS


Digital watermarking has been proposed as a way to claim protection. In this paper, we try to use the adaptive multi-frequency of wavelet transformation, and the human visual system to improve the traditional watermarking. We propose an adaptive watermarking generation technique by embedding watermark in all sub-bands. It makes a watermark to have a coexisting property in part of space domain and frequency domain of protected image. The HVS (Human Visual System) model will be presented to investigate perceptually characteristic parts of human (e.g. intensity or hue) and how to insert the watermark into the original data, where the perceptual sensitivity of human is relatively low. An evaluation function for similarity is used as an adaptive detection parameter. In our experiments, the results show that the robustness in this watermark generation can protect from wide-range of common attacks. We also propose a new observation for robustness of watermark technique.

Important Links:

Go Back