Portal de Periódicos da CAPES

Variational denoising of diffusion weighted MRI

2009; American Institute of Mathematical Sciences; Volume: 3; Issue: 4 Linguagem: Inglês

10.3934/ipi.2009.3.625

1930-8345

Tim McGraw, Baba C. Vemuri, Evren Özarslan, Yunmei Chen, Thomas H. Mareci,

MRI in cancer diagnosis

In this paper, we present a novel variational formulation forrestoring high angular resolution diffusion imaging (HARDI) data. Therestoration formulation involves smoothing signal measurements overthe spherical domain and across the 3D image lattice. Theregularization across the lattice is achieved using a totalvariation (TV) norm based scheme, while the finite element method(FEM) was employed to smooth the data on the sphere at each latticepoint using first and second order smoothness constraints. Examplesare presented to show the performance of theHARDI data restoration scheme and its effect on fiber directioncomputation on synthetic data, as well as on real data setscollected from excised rat brain and spinal cord.