基于核局部线性嵌入算法的图像去噪方法

利用局部线性嵌入算法进行图像去噪时,如果局部近邻样本呈现非线性关系,图像去噪效果会受到影响。针对该问题,提出基于核局部线性嵌入算法的图像去噪方法。通过非线性核函数将样本映射到高维线性空间,在高维空间运用局部线性嵌入算法进行图像去噪。实验结果表明,该方法能有效地对高维非线性图像进行去噪,性能优于中值滤波算法和局部线性嵌入算法。

Image Denoising Method Based on Kernel Locally Linear Embedding Algorithm

Locally Linear Embedding(LLE) algorithm can be used to solve image denoising problem, but when the nearest neighbor samples are nonlinear, the performance of image denoising is degraded. Aiming at this problem, this paper uses Kernel Locally Linear Embedding(KLLE) algorithm to solve image denoising problem in this paper. Image samples are mapped by means of nonlinear kernel function to high dimensional feature space, KLLE algorithm is used to solve image denoising problem in the high dimensional space, which is effective for high-dimensional non-linear image denoising problem. Experimental results show that proposed method is superior to LLE algorithm and median filtering.