摘要
主要研究了低水印嵌入强度下数字图像水印的检测方法 ,提出了一个在 DCT域内应用径向基函数 (RBF)神经网络检测数字图像水印的方案。该方案通过设计一个基于最小二乘学习算法的 RBF神经网络 ,利用它良好的非线性映射能力从待测水印图像中提取出嵌入的水印 ,从而实现对数字图像水印的有效检测。与直接基于相关技术的数字图像水印检测方案相比 。
A new scheme for detecting digital image watermarks in DCT domain is presented. It is based on the radial basis function(RBF) of neural network with an RLS algorithm. By using its good nonlinear mapping characteristics, the digital image watermarks can be effectively detected after extracting the embedded watermark from the watermarked image. Compared to the schemes based on the direct cross correlation technology, the scheme has an excellent detecting ability under the condition of relatively low level embedded watermark intensity.
出处
《数据采集与处理》
CSCD
2001年第4期498-503,共6页
Journal of Data Acquisition and Processing
