Skip to Main content Skip to Navigation
Conference papers

Adaptive Multiresolution Denoising Filter for 3D MR Images

Abstract : In MR imaging, denoising is an important issue in some clinical uses. Recently a new filter has been introduced by Buades: the Non-Local (NL-) means filter. Based on the natural redundancy of the images, this filter uses comparison of image patches in order to achieve the denoising. Due to its simplicity and high denoising performance, this filter is one of the best state-of-the-art methods. Nevertheless, according to the nature of the underlying structures the filtering parameters should be adapted. To implicitly optimize the amount of denoising according to the underlying structure, an adaptive multiresolution framework is proposed.
Complete list of metadatas
Contributor : Pierrick Coupé <>
Submitted on : Tuesday, November 29, 2011 - 6:04:12 PM
Last modification on : Thursday, March 12, 2020 - 3:06:35 PM
Document(s) archivé(s) le : Monday, December 5, 2016 - 9:36:29 AM


Files produced by the author(s)



Pierrick Coupé, José Manjón, Montserrat Robles, Louis Collins. Adaptive Multiresolution Denoising Filter for 3D MR Images. Organization for Human Brain Mapping 2009 Annual Meeting, Jun 2009, United States. pp.S50, ⟨10.1016/S1053-8119(09)70119-3⟩. ⟨hal-00645467⟩



Record views


Files downloads