[BIC-announce] Fwd: PhD Oral Defense - Peter Savadjiev

[Dr. Bruce Pike]

>
> MCGILL UNIVERSITY
> SCHOOL OF COMPUTER SCIENCE
> Ph.D. Oral Defense
>
>
> SPEAKER: 	Peter Savadjiev
> TITLE: 	Perceptual organization in diffusion MRI: curves and  
> streamline flows
> DATE: 	Thursday October 9, 2008
> TIME:		10:30 a.m.
> PLACE:	McConnell Eng. Bldg., Room 437
>
> ABSTRACT:
>
> This thesis proposes a new computational framework for the modeling of
> biological tissue structure in diffusion MRI data. By measuring the
> local Brownian motion of water molecules, diffusion MRI provides
> estimates of local fibre orientations in tissues such as the white
> matter of the brain. Over the last years, diffusion MRI has become an
> important tool for the in vivo study of brain connectivity.
> Nevertheless, the inference of the structure of white matter fibres is
> still an open problem.
>
> The methodology introduced in this thesis is based on differential
> geometry and perceptual organisation. The key ideas are to model white
> matter fibres as 3D space curves, to view diffusion MRI data as
> providing information about the tangent vectors of these curves, and  
> to
> frame the problem as that of inferring 3D curve geometry from a
> discretized, incomplete, and potentially blurred and noisy field of
> tangent measurements. Inspired by notions used in perceptual
> organisation in computer vision, we develop local geometrical
> constraints which guide the inference process and ultimately result in
> the recovery of the underlying fibre geometry.
>
> We start by introducing a notion of co-helicity between triplets of
> orientation estimates, which is incorporated in a geometrical  
> inference
> process. This process is referred to as 3D curve inference, and it
> estimates the parameters of the local best-fit osculating helix for  
> each
> orientation in the dataset. Based on this, a relaxation labeling
> framework is set-up for the regularization of diffusion MRI data.
>
> We then develop a method based on 3D curve inference for the
> identification of complex sub-voxel fibre configurations in high  
> angular
> resolution diffusion (HARD) MRI. Based on the inferred osculating
> helices for each ODF maximum in the data volume, the method labels
> voxels in one of three categories: (1) voxels with a single (possibly
> curving) fibre, (2) voxels with a fanning fibre configuration, and (3)
> voxels with a crossing fibre configuration.
> The focus then shifts from modeling individual curves to the  
> modeling of
> sets of dense, locally parallel 3D space curves which we refer to as
> streamline flows. We develop a differential geometric characterization
> of such structures by considering the local behaviour of the  
> associated
> 3D frame field, leading to the associated tangential, normal and
> bi-normal curvature functions.
> Based on minimal surface theory, we adopt a generalized helicoid model
> as an osculating object and develop the connection between its
> parameters and these curvature functions. This osculating object is
> incorporated in a regularization scheme using relaxation labeling.  
> It is
> also incorporated in a white matter tract segmentation method based on
> geometrical consistency between inferred osculating objects.
>
> The algorithms in this thesis are validated experimentally through
> regularization and tractography studies with in vivo diffusion tensor
> (DT) and HARD data from a human brain, as well as data from a  
> biological
> phantom and synthetic data, for the case of the 3D curve inference
> method.
>
> Committee:	
> Prof. J. Pineau (Chair)		Prof. K. Siddiqi (Supervisor)
> Pro-Dean (TBA)
> Prof. M. Langer			Prof. F. Ferrie
> Prof. G.B. Pike (External member)	
>
>
> ALL ARE WELCOME
>