基于低秩非局部稀疏表示的图像去噪模型  

Image denoising model based on low rank and nonlocal sparse representation

作　　者：王宏宇[1] 陈冬梅[1] 王慧[2] 

机构地区：[1]辽宁行政学院信息中心,辽宁沈阳110161 [2]沈阳农业大学信息与电气工程学院,辽宁沈阳110866

出　　处：《燕山大学学报》2017年第3期272-277,共6页Journal of Yanshan University

基　　金：国家自然科学基金资助项目(61673281)

摘　　要：在去除加性高斯白噪声的过程中,为克服图像失真,提高图像视觉质量,使图像之间联系更加密切,本文提出了一种基于低秩非局部稀疏表示的去噪算法模型。在该模型中,首先通过PCA方法线训练字典得到稀疏字典集,然后用奇异值分解求解低秩问题,最后用欧拉-拉格朗日方法得到去噪图像的能量泛函,从而重构图像。仿真实验结果表明,提出的新算法不仅能克服图像失真,改善图像视觉质量,还提高了峰值信噪比和图像相似度。For removal additive white Gaussian noise,in order to overcome distortion of the image and improve the visual quality of iamge,especially making close between the information of the image,a denoising algorithm based on low rank and sparse representation is proposed in this paper.In this model,first it is using PCA ways to train online dictionary that we can get sparse dictionary set,and then it is used to solve the low rank problem with singular value decomposition.Finally,it is using the Euler Lengrand way to the denoising image for energy function,so we can reconstruct denoised image.A lot of experiments show that the proposed algorithm can not only overcome the image distortion and improve quality of image,but also have a high peak signal to noise ratio and image similarity.

关 键 词：非局部相似 稀疏表示 高斯噪声 低秩 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]