基于支持向量机的NSCT域自适应图像水印算法

为进一步提升支持向量机水印算法鲁棒性，提出基于支持向量机的NSCT域自适应图像水印算法。主要思想是根据图像自身特征生成自适应嵌入水印序列，利用模糊核聚类和支持向量机对NSCT低频系数进行分类，选取适合嵌入水印的低频系数，然后利用支持向量机建立NSCT邻域系数的关系模型，自适应完成水印嵌入。算法具有良好的不可感知性、安全性，并通过嵌入自适应水印达到全盲水印检测。实验结果表明，提出算法对高斯噪声、椒盐噪声、低通滤波、中值滤波、均值滤波、JPEG、旋转、平移和尺寸缩放有很强的鲁棒性。

Adaptive Image Watermarking Algorithm for NSCT Domain Based on Support Vector Machine

In order to further improved the robustness of support vector machine watermarking algorithm, proposed an adaptive image watermarking algorithm for NSCT domain based on support vector machine.The main idea was to generate the adaptive embedding watermarking sequence according to the characteristics of the image itself. Used Fuzzy kernel clustering and support vector machines to classify the NSCT low-frequency coefficients.Selected the low frequency coefficients suitable for embedding the watermark.Then, it established the NSCT neighborhood coefficient relation model by using the support vector machine.The algorithm had good imperceptibility and security, and achieves full blind watermarking by embedding adaptive watermarking sequences.The experimental results show that the proposed algorithm has strong robustness to Gaussian noise, salt and pepper noise, low pass filtering, median filtering, mean filtering, JPEG, rotation, translation and size scaling.