Image Processing and Vision Seminar

 

Time:

Mondays 4:00 pm - 5:00 pm.

Location:

IPAM 1200.

Scope:

Invited speakers and graduate students give informal presentations from their research. Topics include image analysis and computer vision techniques by partial differential equations, wavelets, statistics, or combined approaches.

Organizers:

Selim Esedoglu (Mathematics Department): esedoglu@math.ucla.edu

Luminita Vese (Mathematics Department) : lvese@math.ucla.edu

Alan Yuille (Statistics and Psychology Departments): yuille@stat.ucla.edu

 

Potential speakers are invited to contact one of the organizers.

 

Graduate students attending all the seminar meetings will receive a one unit credit.

 

Schedule Fall 2004

Monday, September 27, 2004: Gerd Teschke, University of Bremen, Germany, "Wavelets and variational principles in image processing".

Monday, October 4, 2004: Yalin Wang, Mathematics UCLA, "Computational Conformal Geometry for Brain Mapping",
Abstract: We developed a general method for global surface conformal parameterization. For genus zero surface, our algorithm can find a conformal mapping between any two genus zero manifolds by minimizing the harmonic energy. We apply this algorithm to the cortical surface matching problem. We further extended the algorithm to find a 3D volumetric harmonic map from a 3D brain volumetric model to a solid sphere. We proposed a new algorithm, sphere carving method, which can build 3D tetrahedral mesh from a set of images with topology preservation. New research on surface matching and 3D face recognition will be discussed and experimental results will be reported.

Monday, October 11, 2004: Roland Wilson, University of Warwick, UK, visiting IPAM, Multiresolution Random Field Models for Image Processing.

Monday, October 18, 2004: Marc Droske, Germany, Level Set techniques for multimodal image registration and implicit formulations for fourth-order geometric diffusion problems.
Abstract: A method for the registration of a pair of images of different modalities is presented. It is based on a variational formulation depending on the image morphology rather than intensity based properties of the images, motivated by the assumption that at least geometric features of the images are similar. In addition to the alignment of morphologies a coupled free discontinuity problem based on the Mumford-Shah functional is used to enhance alignment of dominant edges. In the second part of the talk an overview of level set based gradient flows of, possibly anisotropic, geometric energies will be given and results be presented.

Monday, October 25, 2004: Talks at IPAM in the MGA.

Monday, November 1st, 2004: Demetri Terzopoulos, Courant Institute, NYU.

Monday, November 8, 2004: William K. Allard, Department of Mathematics, Duke University. NOTE ROOM: Boelter 4413. On the regularity of level sets of minimizers for denoising models using total variation regularization.

Monday, November 15, 2004: Talks at IPAM in the MGA program

Monday, November 22, 2004: Peyman P. Moghaddam, UBC, "Optimization techniques for image enhancement and practical implementation".
Abstract: I will give a brief review of different optimization techniques following some examples. I will talk about Linear programming and interior point method and nonlinear optimization methods such as Augmented Lagrangian and Log Barrier methods. I will show how practically these problems can be solved.

Monday, November 29, 2004: Xavier Bresson, EPFL, visiting UCLA.

Monday, December 6, 2004: Selim Esedoglu, Dept. of Mathematics, UCLA.

Monday, December 13, 2004: David Strong, Pepperdine University.

Monday, January 10, 2005: Jean-Francois Aujol, visiting UCLA (tentative).