På svenska

Presentation Information     2009-11-09 (15:15)   •  The seminar room at Vi2

Speaker Ewert Bengtsson
Title Biomedical image analysis in many dimensions
Abstract In an invited presentation at the Australian image analysis conference DICTA I will present an overview of our research with the following abstract. In this CBA seminar I will give a preview of that presentation both to give you an overview of how I see our current work and to allow you to correct my worst misunderstandings of it:

Imaging methods provides rapidly improving and indispensible tools for modern biological and medical research. The images span spatial dimensions from nanometers to meters and can be two dimensional, three dimensional, four dimensional (3D+time) or even five dimensional (with an added spectral dimension). To fully utilize the many new types of images we need to be able to extract quantitative information and/or process many thousand images to obtain sufficient statistics. This requires image analysis methods that are suitable for the different tasks.

At the Centre for Image Analysis in Uppsala, Sweden our mission is to "develop better methods, algorithms and systems for applications within primarily biomedicine, forestry and the environmental sciences". We are currently working with images ranging from electron microscopy of proteins and viruses to tomography of body organs. Many of our methods are based on discrete geometry approaches where we are active also on a fundamental theoretical level but we also borrow ideas from literature and colleagues elsewhere. In this presentation I will provide some illustrative examples from our work, highlight some of the methods and results we have obtained and make some general comments about the challenges and opportunities of the field. Even though the presentation does not have the ambition of providing a comprehensive review of the whole biomedical image analysis field it will hopefully illustrate some current problems and opportunities in this exciting part of science.