Talk – Professor Ela Claridge
On March 25, on the occasion of the PhD defense by Alize Scheenstra, Professor Ela Claridge – PhD, Medical Imaging Lab, School of Computer Science, University of Birmingham, U.K. – will give a talk.
Title: Why we use physics-based models of image formation for the analysis of multispectral images of human tissues
Date and time: March 25, 10:00 – 11:00 am
Place: Collegezaal 4, LUMC
Abstract: Image analysis typically deals with patterns and features derived solely from image data. By considering the physics of image formation we can extend image interpretation to the analysis of properties of the biological tissues, and not only of images of the tissues. Light which enters a tissue interacts with its components, and through these interactions (mainly absorption and scatter) the spectral composition of light is altered in a characteristic way. The remitted light therefore bears an imprint of the tissue properties. By modelling this process for the entire range of parameters that characterise the optical and structural properties of a given tissue we build an imaging model of the tissue which captures the relationship between the tissue properties and the remitted spectra. This model can then be used in a number of ways, including:
- selection of optimal filters for multispectral imaging;
- training of statistical models when the ground truth for a tissue is difficult to obtain;
- quantitative estimation of histological parameters characterising a tissue.
The applications of these ideas will be shown for a number of tissues including the skin, the eye and the colon.
Biography: Professor Ela Claridge is a Professor of Medical Image Analysis at the Medical Imaging Lab, School of Computer Science, University of Birmingham, U.K. Her research area is image understanding and computer vision, especially in application to medical images. This research encompasses a number of different disciplines including physics of image formation, human visual perception, process of medical diagnosis and evolutionary computation for computer vision. Results of her research are being used in clinical practice to help with diagnosis.
