Research Training Group 2154 - Materials for Brain

RTG Online Colloquium talk by Dr. Christian Lucas: Deep Learning for the prediction of lesion progression in ischemic stroke using multivariate CT data

Institute of Materials Research, Biological Characterization, Helmholtz-Zentrum Geesthacht

Oct 15, 2020 from 05:00 PM to 06:00 PM


Link to the video-meeting
No access code required.

Christian Lucas

Cerebrovascular diseases, in particular ischemic stroke, are one of the leading global causes of death in developed countries. Perfusion CT and/or MRI are ideal imaging modalities for characterizing affected is-chemic tissue in the hyper-acute phase. If infarct growth over time could be predicted accurately from func-tional acute imaging protocols together with advanced machine-learning based image analysis, the ex-pected benefits of treatment options could be better weighted against potential risks. The quality of the out-come prediction by convolutional neural networks (CNNs) is so far limited, which indicates that even highly complex deep learning algorithms are not fully capable of directly learning physiological principles of lesion tissue progression through weak supervision due to a lack of data (e.g., follow-up segmentation). In this talk, I am going to give an overview about current Deep Learning methods in biomedical image analysis to pre-dict tissue outcomes in acute stroke and simulate lesion progression.

Add to your iCal calendar