Non-local MRI denoising using random sampling

被引:16
作者
Hu, Jinrong [1 ,2 ]
Zhou, Jiliu [2 ]
Wu, Xi [2 ]
机构
[1] Xihua Univ, Sch Comp & Soft Engn, Chengdu 610039, Peoples R China
[2] Chengdu Univ Informat Technol, Dept Comp Sci, 24 Block 1,Xuefu Rd, Chengdu 610225, Peoples R China
基金
美国国家科学基金会;
关键词
MRI; Denoising; Non-local means; Random sampling; Sampling strategy; KERNEL REGRESSION; IMAGE; NOISE; FILTER; SCALE;
D O I
10.1016/j.mri.2016.04.008
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 ; 100207 ; 1009 ;
摘要
In this paper, we propose a random sampling non-local mean (SNLM) algorithm to eliminate noise in 3D MRI datasets. Non-local means (NLM) algorithms have been implemented efficiently for MRI denoising, but are always limited by high computational complexity. Compared to conventional methods, which raster through the entire search window when computing similarity weights, the proposed SNLM algorithm randomly selects a small subset of voxels which dramatically decreases the computational burden, together with competitive denoising result. Moreover, structure tensor which encapsulates high-order information was introduced as an optimal sampling pattern for further improvement. Numerical experiments demonstrated that the proposed SNLM method can get a good balance between denoising quality and computation efficiency. At a relative sampling ratio (i.e. xi = 0.05), SNLM can remove noise as effectively as full NLM, meanwhile the running time can be reduced to 1/20 of NLM's. (C) 2016 Elsevier Inc. All rights reserved.
引用
收藏
页码:990 / 999
页数:10
相关论文
共 41 条
  • [11] Pointwise shape-adaptive DCT for high-quality denoising and deblocking of grayscale and color images
    Foi, Alessandro
    Katkovnik, Vladimir
    Egiazarian, Karen
    [J]. IEEE TRANSACTIONS ON IMAGE PROCESSING, 2007, 16 (05) : 1395 - 1411
  • [12] NONLINEAR ANISOTROPIC FILTERING OF MRI DATA
    GERIG, G
    KUBLER, O
    KIKINIS, R
    JOLESZ, FA
    [J]. IEEE TRANSACTIONS ON MEDICAL IMAGING, 1992, 11 (02) : 221 - 232
  • [13] A Nonlocal Maximum Likelihood Estimation Method for Rician Noise Reduction in MR Images
    He, Lili
    Greenshields, Ian R.
    [J]. IEEE TRANSACTIONS ON MEDICAL IMAGING, 2009, 28 (02) : 165 - 172
  • [14] From Local Kernel to Nonlocal Multiple-Model Image Denoising
    Katkovnik, Vladimir
    Foi, Alessandro
    Egiazarian, Karen
    Astola, Jaakko
    [J]. INTERNATIONAL JOURNAL OF COMPUTER VISION, 2010, 86 (01) : 1 - 32
  • [15] Noise-Driven Anisotropic Diffusion Filtering of MRI
    Krissian, Karl
    Aja-Fernandez, Santiago
    [J]. IEEE TRANSACTIONS ON IMAGE PROCESSING, 2009, 18 (10) : 2265 - 2274
  • [16] Kwan R.K.-S., 2010, BRAINWEB SIM BRAIN D
  • [17] Kernel regression based feature extraction for 3D MR image denoising
    Lopez-Rubio, Ezequiel
    Nieves Florentin-Nunez, Maria
    [J]. MEDICAL IMAGE ANALYSIS, 2011, 15 (04) : 498 - 513
  • [18] Nonlocal Transform-Domain Filter for Volumetric Data Denoising and Reconstruction
    Maggioni, Matteo
    Katkovnik, Vladimir
    Egiazarian, Karen
    Foi, Alessandro
    [J]. IEEE TRANSACTIONS ON IMAGE PROCESSING, 2013, 22 (01) : 119 - 133
  • [19] MRI denoising using Non-Local Means
    Manjon, Jose V.
    Carbonell-Caballero, Jose
    Lull, Juan J.
    Garcia-Marti, Gracian
    Marti-Bonmati, Luis
    Robles, Montserrat
    [J]. MEDICAL IMAGE ANALYSIS, 2008, 12 (04) : 514 - 523
  • [20] MRI noise estimation and denoising using non-local PCA
    Manjon, Jose V.
    Coupe, Pierrick
    Buades, Antonio
    [J]. MEDICAL IMAGE ANALYSIS, 2015, 22 (01) : 35 - 47