MedTech West seminar – Parametric models for brain segmentation – not dead just yet

In recent years, non-parametric, multi-atlas techniques have become very popular for automated segmentation of medical images. In this talk, I will try to make a case instead for old-fashioned, parametric models in brain MR segmentation. In particular, I will argue that these methods can combine state-of-the-art parcellation performance with a unique set of advantages, including high computational speed, automatic adaptiveness to changes in image contrast when different scanner platforms and pulse sequences are used, and the ability to handle the multi-contrast (vector-valued intensities) MR data acquired in clinical settings.

The seminar will be presented by Koen Van Leemput, Ph.D. at Athinoula A. Martinos Center for Biomedical Imaging, Harvard Medical School, USA and Department of Applied Mathematics & Computer Science, Technical University of Denmark.

Date: February 11, 2014
Time: 11:30-12:15 (-13:00 incl lunch)
Room: Förmaket, Vita Stråket 12, Sahlgrenska University Hospital

Register by sending an email to Please write “Seminar February 11” in the subject line. Last day for registration is February 7.