Ingrid Daubechies, Duke University (US)
Sparsity in Signal Analysis and in Computation
In a wide range of applications, we seek to capture efficiently the most salient characteristics of a signal or phenomenon. In other words, we seek a presentation of a family of functions or objects that manages to be both accurate (with respect to the particular aspects in which we are interested) and succinct -- that is, a *sparse* representation. Such sparse representations have many advantages for analysis, storage or computation. In some cases, we have sufficient insight in the mathematical properties of what we are studying to "guess" ways of obtaining sparse representations -- with hindsight, the use of wavelets for images can be viewed as a particular instance. In other cases, we try to use mathematical tools to guide us to the construction of such a sparse representation; very exciting current developments in this direction bring differential geometry and topology to bear on applications in ways that would probably have surprised mathematicians of past generations.
The presentation will give an overview of work by others as well as by the speaker.
Short Bio: here
Konrad Polthier, Freie Universität Berlin (DE)
Covering and Uncovering Surfaces
Multivalued functions and differential forms naturally lead to the
concept of covering surfaces and more generally of covering manifolds.
This talk will review, discretize and illustrate basic concepts of coverings
and discuss their appearance in recent geometry processing algorithms
including surface and volume parameterization. We will also embark on
the experimental path using scissor and paper as well as the zometool kit.
Konrad Polthier is MATHEON professor of mathematics at Freie Universit‰t Berlin
and co-chair of the Berlin Mathematical School. Polthier received his PhD from
University of Bonn in 1994, and headed research groups a Technische Universit‰t Berlin
and Zuse Institute Berlin before joining FU Berlin. His research interests are in
discrete differential geometry, applied geometry and mathematical visualization.
Results of his research have found industrial applications in computer graphics,
computer aided design and architecture.
Dr. Polthier co-edited several books on mathematical visualization, and co-produced educational
video films. His video MESH ñ A Journey through Discrete Geometry (www.mesh-film.de, joint
with Beau Janzen, Los Angeles) received international awards including ìBest Animationî at
the New York International Independent Film Festival. Polthier served as paper co-chair on
international conferences including ACM/Eurographics Symposium on Geometry Processing.
For more details see: http://www.polthier.info
Niloy Mitra, University College London (UK)
In Search of Geometric Simplicity
As 3D model repositories continue to grow, there is a pressing need to develop tools to analyze, explore, organize, and synthesize such massive model collections. In this talk, we will talk about our recent efforts in this direction both in terms of analysis and structure-aware synthesis. We will demonstrate how such discovered structures play an important role in different applications including image-based scene understanding, geometry processing, and fabrication-aware form finding. We will explore correlations between form and function of objects, and discuss some of the key challenges we are facing in this area.
Niloy J. Mitra is an associate professor (Reader) in the Department
of Computer Science at University College London (UCL). Earlier, he co-founded the Geometric Modeling and Visualization Center (GMSV) center at KAUST. He received his Ph.D. degree and Masters in Electrical Engineering from Stanford University. His research interests include shape understanding, fabrication-aware design, geometric modeling, recreational art, and computer graphics in general.
Lihi Zelnik-Manor, Dept. of Electrical Engineering, Technion (Israel)
Finding the Important Pixels
In recent years more and more cameras record the world using more and more pixels. Watching and processing all this data takes lots of time, which we don’t want to spend. But do we really need all the pixels? the answer is no. When multiple cameras view the same scene often a single “good” view suffices to visualize what’s going on. Within a single view keeping only the “important” pixels suffices to convey the story the image/video tells. Finding these salient pixels in an image or a video is useful for a broad range of tasks. Applications such as cropping, compression, and editing all aim at preserving the important regions intact.
Therefore, in this talk I will explore the state-of-the-art methods for finding the salient pixels in both images and video.
Lihi Zelnik-Manor is an assistant professor in the Electrical Engineering Department at the Technion in Israel. Before that she worked as a post-doctoral fellow in the California Institute of Technology (Caltech). She holds a PhD and MSc (with honors) in Computer Science from the Weizmann Institute of Science in Israel.
Dr. Zelnik-Manor’s awards and honors include the Israeli high-education planning and budgeting committee scholarship, the Sloan-Swartz postdoctoral fellowship, the best Student Paper Award at the IEEE SMI'05, the AIM@SHAPE Best Paper Award 2005 and the Gutwirth prize for the promotion of research.