A Multiwavelet based digital watermarking algorithm using texture measures

Abstract

Digital watermarking technique is a leading international research field in recent years. Though watermarking algorithms based on scalar wavelet transforms have become a hot-discussed research aspect and fruitful achievements have been obtained. but the ones based on multi-wavelet transforms have appeared few. Multi-wavelet has symmetric, orthogonality and compact support at the same time. They are important properties in image processing. In the paper, watermarking system in multi-wavelet domain has been studied, and a multi-wavelet transform based JND digital watermarking algorithm has been proposed. By calculating the textutre measures of the multi-wavelet coefficients in the low frequency, the JND calculation model has been improved. The experiment result shows that by using JND calculating model based on texture measures, the robustness of the JND threshold is improved, and the watermark embeded position can be located exactly, at the same time the digital watermark extraction result is significantly improved.