Open AccessCombining interpolation and 3D level set method (I+3DLSM) for medical image segmentation
Author(s) -
Doshi T.,
Di Caterina G.,
Soraghan J.,
Petropoulakis L.,
Grose D.,
MacKenzie K.,
Wilson C.
Publication year - 2016
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
ISSN - 1350-911X
DOI - 10.1049/el.2015.3959
Subject(s) - interpolation (computer graphics) , segmentation , artificial intelligence , computer vision , trilinear interpolation , image scaling , multivariate interpolation , computer science , nearest neighbor interpolation , stairstep interpolation , image segmentation , pattern recognition (psychology) , bilinear interpolation , mathematics , image (mathematics) , image processing
A combined interpolation – three‐dimensional level set method (3D LSM)‐based segmentation process is presented. The performance in terms of accuracy of the 3D LSM in the segmentation of throat regions from highly anisotropic magnetic resonance imaging (MRI) volumes, with and without an interpolation step is evaluated. Qualitative and quantitative results from real MRI data suggest that performing interpolation, to reconstruct isotropic MRI volumes, prior to 3D LSM improves the accuracy of the segmentation results, compared to interpolation post 3D LSM and no interpolation at all.