18. – 23. März 2007, Dagstuhl Seminar 07121
Experimental Fluid Mechanics, Computer Vision & Pattern Recognition
Die Dagstuhl-Stiftung erhielt eine Spende von:
|•||LaVision - The Imaging Company|
Auskunft zu diesem Dagstuhl Seminar erteilt
Recent advances in imaging and measurement techniques have enabled the generationof huge amounts of data in the field of Experimental Fluid Mechanics, providinga unique basis for the understanding of spatial structures in unsteady flows. Progressin this field has an immediate impact on different areas ranging from aerodynamics tobiology, and related industrial applications.
The design of computational approaches for the evaluation of image data of fluidsraises a range of unsolved problems. A non exhaustive list of these problems includes multiscale image representation, image motion computation, image sequence segmentation, representation of physical prior knowledge, probabilistic inference, spatiotemporal event recognition, and stochastic models for tracking. These fundamental issues are investigated in Computer Science (Computer Vision, Pattern Recognition) as well, although in connection with different applications areas. Despite a confluence of methodological interests, there has been surprisingly only little interaction between Computer Science and Experimental Fluid Mechanics, so far.
In view of the scientific importance, two research programmes have been established quite recently, bringing together researchers from Fluid Mechanics and Computer Science: the national German priority programme on Image Measurements in Fluid Mechanics (DFG-SPP 1147), and the European project Fluid Image Analysis and Description.
Objective of the Dagstuhl seminar is to address these issues in a larger international context with leading experts from Fluid Mechanics and Computer Science.
2. Topics to be discussed are:
- Image Measurements, Image Processing
- Computer Vision and Spatio-Temporal Pattern Recognition
- Synergy of Experiments and Computer Simulations in Fluid Mechanics
The Dagstuhl seminar will provide an ideal platform for communicating approaches and ideas between researchers of two scientific communities that pursue similar overall goals: the design of computational systems for the understanding of complex spatiotemporal phenomena from image measurements.
Attendees will be asked specifically to report not only on past experience and present state-of-the-art but to devote a significant portion of their presentation to future needs and capabilities. On goal of the Seminar is to develop a realistic vision of the future, including identifiable initial steps. Novel contacts and cooperation may be expected that will lead to joint research work with a high scientific impact.
Experiments in Fluids will publish a Special Issue consisting of papers selected from those contributed to the Seminar. To maximise the possibility of being selected, authors are encouraged to prepare their papers with a quality level consistent with being submitted directly to a journal for review.
- Interdisciplinar Experimental Fluid Mechanics
- Computer Vision
- Image Processing
- Pattern Recognition
- Modelling & Simulation
- Particle image velocimetry
- Advanced measurements and imaging
- Coherent structures
- Turbulent flows
- Flow control
- Multiscale image representations
- Variational modeling
- Pde-based image processing
- Optical flow
- Image registration
- Image segmentation
- Karhunen-Loeve transformation
- Feature selection
- Probabilistic modeling and inference
- Spatio-temporal tracking
- Event recognition
- Sequential Monte-Carlo methods
- Direct numerical simulation
- Reduced order modeling
- Statistical dynamics