BBB Seminar: Arvid Lundervold
Biomedical image analysis - challenges, solutions and local organization
Biomedical image analysis is a multidisciplinary field between science and information technology. It applies digital image processing, computer vision, pattern recognition techniques and mathematical modeling to address quantitative problems in biology and medicine that are linked to recorded signals (images) in space and time.
Biomedical image analysis focuses on post-acquisition challenges such as object detection and object tracking; quantification and analysis of size, shape, texture, and motion; model-based estimation of physiological and biophysical parameters such as diffusion, flow, and permeability; spatial alignment and integration of information from different modalities and imaging devices, as well as pre-processing steps related to spatiotemporal filtering, noise suppression and image restoration.
During the last decade biomedical imaging equipment and imaging techniques (e.g. MRI, PET/CT, ultrasound, confocal microscopy, MALDI MS imaging; diffusion tensor imaging, multispectral imaging, tissue velocity imaging, STED, FRET & FRAP, QD and single particle tracking) have attained a high level of sophistication and have also been more available. Imaging in biology and medicine is now performed at high-throughput and at a large range in both spatial and temporal scales, where recorded data during one experiment or patient examination can be at hundreds of megabytes and at high dimensionality (3-D+time, even 5-D). It is thus very difficult and time consuming, in some cases even not possible, to comprehend or disentangle information in such digital data sets without automation and the use of proper analysis tools.
During the talk I will point to challenges and future perspectives for biomedical image analysis in general, and also to specific solutions and methodological approaches from our local research regarding brain imaging and fiber tracking, functional kidney imaging, and automated detection of tunneling nanotubes (TNTs) in 3-D microscopic images. Finally, the organization of biomedical image analysis in Bergen, in terms of research networks and training (i.e. the MedViz initiative, the BBG/Vis group, the new Master program track in Biomedical Image Sciences), and the NRC-funded Norwegian Interdisciplinary Research Training Program in Medical Imaging, anchored at NTNU, will be presented.
Chair: Jaakko Saraste , Department of Biomedicine