| Speaker | | Görkem Saygılı, PhD |
| Date | | February 8th, 2019(Friday) |
| Time | | 10.00-11:00 |
| Location | | Informatics Institute, Room 411
|
Abstract
Predicting a confidence score/error is an important task in many sub-fields of computer vision such as stereo matching and image registration. This task becomes even more challenging in medical image registration since available ground truth is very limited. Our recent findings show that the registration error can be predicted with high accuracy via using machine learning tools together with modality independent neighbour descriptors, the variation of deformation fields, and block-matching based features.
Short bio
Görkem Saygılı is an Assistant Professor in the Biomedical Engineering department of Ankara University. He received his B.Sc. degree from Bilkent University Electrical and Electronics Engineering department in 2007, M. Sc. degree from Koç University in 2010 and his Ph.D. degree from Technische Universiteit Delft (TU Delft) in 2015. Before joining Ankara University, he was a Postdoctoral Researcher at Leiden University Medical Center (LUMC) and participated in Human Brain Project (HBP).
Confidence Estimation for Stereo Matching and Medical Image Registration
