基于光滑支持向量机的图像水印技术

Image watermark scheme based on smooth support vector machine

一些基于标准支持向量机的图像水印技术,使图像水印的效果得到一定改善,但降低了图像水印技术的效率。针对这个问题,用光滑支持向量机取代标准支持向量机,结合图像的局部相关特性来确定图像的最佳嵌入位置和嵌入强度,提出一种基于光滑支持向量机的图像水印技术,并做了仿真实验。实验表明,这种技术与以往基于标准支持向量机的图像水印技术相比,不仅效果更优,而且效率也显著提高。

Using a class of new smoothing functions, the problem of smooth support vector machine （SVM） was studied. A new model of SVM, 3rd-order polynomial smooth support vector machine （3SSVM）, was proposed, and its global convergence was established. Numerical experiments were carried out to evaluate 3SSVM, using Newton-Armijo algorithm. The results show that 3SSVM is better than PSSVM and SSVM in the classification performance and computational speed. A better theoretical support is provided for applications of smooth support vector machine.