A Novel Threshold Segmentation Embedding Algorithm Of Digital Watermarking

This article presents a new algorithm of digital watermarking based on fuzzy threshold segmentation and the best permutation which also adaptively adjusts the intensity of embed watermarking determined by the characteristics of the human visual system.They each has its weak point either we embed the watermark into the low frequency bands or the high frequency bands, cause it has better robustness and worse invisibility in the low frequency band by contrary to the high frequency band. Consequently, it is a key question in the watermark technology how to keep balance between the robustness and invisibility of the watermark. Based on fuzzy threshold segmentation we divide the watermarking into two partition of different intension in order to adapt to the invisibility and robustness characterized by the low and medium frequency bands.What's more, permuted measure is introduced to evaluate the permutation of an image quantitatively. The best permutation is achieved by controlling the parameters of chaotic mapping. The image watermarking is permuted to reduce the relativity of original pixels, so the error bits of the extracted watermarking are dispersed as well. Therefore the resistance to crop attack is improved significantly. Chaotic systems have attracted extensive interests in cryptanalysis and communications because of its noise-like, unpredictable yet deterministic orbits and sensitivity to initial condition and parameters into image permutation, and then the permuted watermarking signal is embedded into wavelet transform coefficients of host image.