摘要： 提出一种基于小波域的多功能数字水印算法，根据人类视觉特性和原始图像的自身内容确定可见水印在低频子带和各高频子带的拉伸系数。对嵌入可见水印图像的低频小波系数进行量化，并将合适的量化步长嵌入不可见水印。实验结果表明，该算法生成的水印图像满足可见水印的要求，且不可见水印对一般的图像处理和攻击具有很好的鲁棒性，可以同时实现版权通知和版权保护。

关键词: 可见水印, 小波变换, 人类视觉特性

Abstract: This paper proposes a multi-function digital watermark algorithm based on Discrete Wavelet Transformation(DWT), where the visible watermark’s scaling factors of low frequency and each high frequency are determined by Human Visual System(HVS) and the host image features. Low frequency coefficients of the image embedded with visible watermark are selected to embed invisible watermark using quantization method by appropriate quantifying step size. Experimental results show that the watermarked image generated by the algorithm meets the requirements of visible watermark and the invisible watermark is robust against normal image processing and attacks. The algorithm can realize copyright informed and copyright protected together.

Key words: visible watermark, wavelet transformation, Human Visual System(HVS)

中图分类号: 

• TP391