从图像的退化离散模型出发，对图像进行逐层分块奇异值分解，并自动选取奇异值重组阶数以减少噪声对估计的影响。
Using the discrete image degradation model, a block-based SVD filter scheme was applied for the image denoising with an automatically determined singular value rank.
利用矩阵的奇异值分解和矩阵分块方法，得到了最小二乘解的一般表达式。
By using the method of matrix singular values decomposition, the general expressions of the least squares solutions are given.
首先对图像进行分块，分别提取每块图像的直方图不变矩，组合起来作为数字水印，然后利用奇异值分解的方法将这些水印嵌入到原图像中。
First we go on blocking to image an extract moment invariant of histogram from the blocking image and combine it as the digital watermarking.
应用推荐