Spring 2010
|
February 11
|
Li Wu |
Title: Polynomials and Differential Equations
|
February 25
|
James Baglama |
Title: Krylov Subspace Methods: How to Solve Large Scale Eigenvalue Problems |
March 11
|
Tim Toolan |
Title: Rank-One Modification of the Singular Value Decomposition |
March 18
|
Tomas Francisco,
Yago Vicente, Chris Allen, and Jean-Yves Herve |
Title: Fitting and Registration of Multi-Representation Models of Knee
Cartilage Surfaces
We will present ongoing work on the representation of surfaces of knee
articular cartilage. Our input consists of 3D meshes obtained by
segmentation of MRI data. Our model combines this mesh information with
implicit and parametric (NURBS) representations of the surface. We use
these joint representations to study the local geometry of the
cartilage's surface. We also register surfaces obtained from yearly
scans of the same patient to study the longitudinal evolution of these
local geometric properties. The toolbox used for this project is fairly
typical for computer vision (visual reconstruction): digital geometry,
computational geometry, multidimensional optimization and solution of
nonlinear equations.
|
April 8
|
Pankaj L Ahire |
Title:Strategies for Explicit Image Detection
via Automated Image Analysis
In this talk I will present the work being done by the Human Image
Detection group at the Digital Forensics Center at URI.
I shall give a brief overview of the contemporary framework of Image
analysis, object detection techniques, learning approaches, and
evaluation metrics. The various approaches we tried will then be
explained in context of these techniques.
As this is an ongoing project, I will conclude by presenting our
current results and ideas for future direction.
|
April 22
|
Jean-Yves Herve |
Title: Detection of cylindrical rods in CT scans of soil samples
I will present work that we did in collaboration with the US EPA. EPA
researchers have developed an ecological criterion to estimate the
quality of coastal and marsh soils using the density of live roots and
worm tubes present in a soil sample. Automating the computation of this
criterion requires detecting straight cylindrical "calibration" rods in
volumetric CT scans of soil samples. Although this sounds like a very
simple problem, it turns out to be a fairly complex (and slow) task. I
will explain why and discuss the solutions that we have developed for
this problem. |
April 29
|
Josh Bacon and Shawn Case |
Title (Shawn Case): Quadrature Rules using Szego Polynomials
Title (Josh Bacon): On Szego Polynomials and System Theory |