Thursday, July 17

10:30 AM-12:30 PM
Terman Auditorium

Nonlinear PDE Methods in Image Processing (Part I of II)

Image processing (including enhancement, restoration, segmentation, etc.) has many potential applications in both the physical and the medical sciences. The problems are often computationally intensive and accurate and efficient methods are needed. Traditional image processing methods rely on two major approaches: transform methods and statistical methods. In recent years, a new nonlinear PDE based approach has become popular, because it offers a systematic way of dealing with images with sharp edges while minimizing associated Gibb's phenomenon's of linear methods. As yet, the PDE methods are still not fully developed and much room for improvement exists, both in modeling aspects and computational efficiency.

In this minisymposium, the speakers will address image enhancement, restoration and segmentation applications for grey scale and vector-valued (e.g. color) images. New models will be presented based on variational and geometric approaches. Numerical techniques addressed will include primal-dual interior point methods and sparse linear algebraic methods.

Organizer: Tony F. Chan
University of California, Los Angeles

10:30 Nonlinear PDE Methods in Image Processing and Some Numerical Issues
Pep Mulet, University of Valencia, Spain
11:00 Vector PDE's in Image Processing: Results and Questions
Guillermo Sapiro, Hewlett-Packard Laboratories
11:30 Images as Embedded Maps and Minimal Surfaces
R. Malladi, R. Kimmel, and N. Sochen, Lawrence Berkeley Laboratory
12:00 Interior Point Methods for Image Enhancement
Rene Carmona and Sifen Zhong, Princeton University

AN97 Homepage | Program Updates|
Registration | Hotel and Dormitory Information | Transportation | Program-at-a-Glance | Program Overview

MMD, 5/30/97