利用最近邻域分类的图像去噪算法

Image Denoising Algorithm Using Nearest-Neighborhood Classification

为了在图像去噪的同时较好地保护图像细节,利用最近邻域的算法将数据进行有效的分类,得到了不同的有意义的封闭邻域,从而对突出的边缘细节信息与非边缘细节信息进行有效地分割,较好地改善了方形邻域固定、模糊边缘细节信息的问题.且利用小波分析之后的系数特征,估计出一个最佳阈值并进行阈值去噪.实验表明,该算法可以得到更好的实验结果.

To keep image details during image denoising,an effective algorithm was proposed to classify the data based on nearest-neighborhood method so as to get various close domains.In this way the salient edge and non-edge information can be divided effectively.This proposed method can improve the original method,which may blur the edge.An optimum threshold was estimated to de-noise the degraded image.The experiments show that the scheme can get better result than others.