Abstract:

A robust image watermarking algorithm based on Human Visual System（HVS） and Discrete Wavelet Transform（DWT） is proposed.Host image is decomposed through the discrete wavelet transform.The JND threshold is calculated according to the human visual system characteristics，and the significant coefficients of wavelet are selected.Watermark image is embedded to the different frequency of the significant coefficients of wavelet.And watermark image is embedded by the rule of the improved binary operation，changing wavelet coefficients part of the decimal to improve the quality of the image after the watermark embedded.The experimental results show that the proposed algorithm has enhanced the robustness effectively under the premise of guaranteeing watermark invisibility，and it is robust against JPEG compression，additive noise，zooming，cut and median filtering attack.

Key words: Human Visual System（HVS）, Discrete Wavelet Transform（DWT）, Just Noticeable Difference（JND） threshold, robustness

摘要： 提出了一种基于人类视觉系统（HVS）和离散小波变换（DWT）的图像鲁棒数字水印新算法。将宿主图像进行小波分解，根据人类视觉系统的掩蔽特性计算出JND视觉门限值，选择重要小波系数，将数字水印分别嵌入到重要小波系数的不同频率域中。水印嵌入采用改进的二值运算方法，修改小波系数的小数部分，以提高水印嵌入后图像的质量。实验证明算法在保证水印不可见性的前提下有效地提高了算法的鲁棒性，对JPEG压缩、图像添加噪声、缩放、剪切和中值滤波等攻击有很强的鲁棒性。

关键词: 人类视觉系统, 离散小波变换, JND门限, 鲁棒性