基于小波包和DCT变换的双水印技术

An Algorithm of Bi-watermarking Based on Wavelet Packet and DCT Transform

目前较流行的水印算法是在图像的时空域或变换域中感知重要的系数嵌入水印。然而感知重要的部分通常是图像的低频部分,当图像经历某些变换操作时,从低频提取的水印会产生严重的质量退化。针对这一问题,该文在结合小波包变换和离散余弦变换的同时,在图像小波包分解后的低频和高频部分选取适当系数分别嵌入同一个水印。水印采用二值图像,在水印嵌入之前对水印图像进行Arnold变换以增强算法的隐蔽性和鲁棒性。实验结果表明,这一方案比单重水印方案能够更有效地抵抗JPEG有损压缩、剪切、滤波和噪声等常见的水印攻击,具有更广泛的适应性和更强的鲁棒性。

The watermark is often embedded in the important perceptual parts of spatial domain or transform domain which are usually low frequency parts of the image. When the image is undergoing some kind of attacks, the watermark extracted from the image will be seriously degraded. In order to solve this problem, the paper presents a so - called Bi - watermarking Algorithm Based on Wavelet Packet and DCT transform. Choosing some parts of appropriate coefficients to insert the same watermark into low frequency and high frequency part of the wavelet packet domain. The watermark is bi - level image, before the watermark is embedded into the original image, we do Arnold transform to increase the concealment and robustness of the Algorithm. Experimental results show that the proposed new scheme has stronger robustness than the Single -watermarking method for common image processing and is adapted to much more situations.