基于提升小波变换和离散余弦变换的彩色图像水印算法

针对一些图像水印算法不可见性与鲁棒性不易平衡的问题,提出一种基于提升小波变换和离散余弦变换相结合的水印算法。首先提取载体图像的边缘特征构造二值水印图像,并使用Arnold变换和Logistic混沌映射对水印图像进行加密。然后,将RGB图像转到YCuCv空间,对Cv通道进行提升小波变换(LWT)和分块离散余弦变换(DCT)。最后将加密的水印嵌入每个DCT变换块的低频系数中。实验结果表明,嵌入水印的图像峰值信噪比(PSNR)值均高于43 dB,进行各种常见攻击后归一化相关系数(NC)值接近1。因此,所提出的方案具有较好的水印不可见性,同时对不同的攻击具有较强的鲁棒性。

Color Image Watermarking Algorithm Based on LWT-DCT

To solve this problem of unbalance between the invisibility and robustness of the image watermarking algorithm, this paper proposes a watermarking algorithm based on lifting wavelet transform and discrete cosine transform. Firstly, extract the edge feature of the cover image to construct a binary watermark image, which was encrypted by using the Arnold transform and the Logistic chaotic map. Then, an RGB image was converted to YCuCv space, the LWT and the block DCT transform were performed on the Cv channel. Finally, the encrypted watermark was embedded into the low frequency coefficients of each DCT transform block. The experimental results show that the PSNR values of the watermarked image are all higher than 43 dB, and the NC values of are close to 1 after various conventional attacks. Therefore, the proposed scheme has better performance in term of invisibility, and strong robustness to different attacks.