Workshop on Geometry and Statistics in Bioimaging: Manifolds and Stratified Spaces

8-12 October 2012 at Sandbjerg Estate, Sønderborg, Denmark

Sunday will be arrival day. The scientific programme starts on Monday morning and ends on Friday around noon.

Scope of the Workshop

This workshop is dedicated to geometric and statistical modeling in biomedical image analysis. The workshop is a result of our desire to bring together researchers from biomedical image analysis, who have an interest in the underlying mathematical constructions, along with mathematicians who have an interest in the underlying practical problems.

Traditional statistical methods analyze data in linear, Euclidean spaces. In many bioimaging problems, however, data is best described in non-Euclidean spaces. Some types of imaging data, such as 2-dimensional shapes, are naturally modeled on manifolds. However, even basic statistical notions such as PCA do not have an obvious well-defined translation into the manifold world. Even more complicated data spaces are needed to describe shapes from 3-dimensional images, which are now standard in bioimaging. Spaces of 3-dimensional shapes are not manifolds, but singular, stratified group quotients, where statistical measurements are not obviously well defined or computable. The class of stratified spaces also contains spaces of graphs and trees, which receive a growing interest in medical image analysis, where graphs and trees actually appear, e.g., as blood vessel networks, airway trees or dendrite trees. The goal of the workshop is to study how geometry affects statistics in these nonlinear data spaces.

Participation

Due to a restricted number of places, this workshop is in principle "by invitation only". However, there are a few free places. For further information, please contact Oddbjørg Wethelund (oddbjorg@imf.au.dk).

Invited Speakers

Scientific Programme Committee

Organizing Committee

Further Information

Please contact Oddbjørg Wethelund, e-mail: oddbjorg@imf.au.dk, Tel +45 8715 5591.