基于噪声分析的混浊数字图像水印算法

本文提出了一种具有较强鲁棒性的混沌数字图像水印算法。首先,根据特定的初始条件和参数产生混沌序列,从而构造出数量大、保密性好、自相关性强、互相关性弱的数宇水印信号。对于水印的嵌入,根据小波分解后高频系数的能量来决定水印嵌入的区域,进而根据各层可容纳的噪声信号强度来控制水印的嵌入强度。水印的检测过程中先对图像进行几何配准,能够有效的提高水印检测的准确性。实验证明该水印嵌入算法易于实现,且具有较强的鲁棒性及视觉上的难以感知性。

A Chaotic Digital Image Watermarking Algorithm Based on Noise Analysis

In this paper, a robust digital image watermarking based on chaotic sequence and noise analysis is proposed. The digital watermark signal is generated through a chaotic sequence, which is determined by initial condition and parameters. The chaotic digital watermark signal has the advantages of massive, high security, strong auto-correlation and weak inter-correlation. The embedding region is chosen according to the energy of high and middle frequency coefficients in wavelet domain. The embedding strength tolerated in each decomposition level is determined by the corresponding noise analysis results. Image geometrical calibration is introduced to improve the performance in watermark detection. Experiments have shown that this proposed algoridim is feasible to implement and the overall performance is enhanced. It provides considerable robustness and imperceptibility.