Harnessing Segmentation Uncertainties for Enhanced Radiotherapy
PhD project
I research uncertainty based segmentation and the benefits of this in the context of radiotherapy.
A bottleneck in implementing adaptive radiotherapy is the segmentation. Both regarding the time to produce it and the extended review needed by physicans. We believe that using uncertainty based radiotherapy will give us a fast segmentation that also reduces the time that physicans need to spend reviewing the segmentation before using it for treatment. In this project we will review different types of uncertainty quantification and how well these translates into finding segmentation errors for deep learning models.
Head of project
Anders GarpebringAssociate professor, combined with clinical employment