October 31, 2024
Computer science doctoral student Basar Demir won the Best Oral Presentation award at the 2024 Workshop for Biomedical Image Registration (WBIR), which was held as part of the 2024 International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI).
The award was given for the presentation of his paper, multiGradICON: A Foundation Model for Multimodal Medical Image Registration.
The paper is a collaborative work authored by computer science doctoral students Demir, Lin Tian, and Thomas Hastings Greer, as well as Roland Kwitt of the University of Salzburg in Austria, Francois-Xavier Vialard of Gustave Eiffel University in France, Raul San Jose Estepar of the Brigham and Women’s Hospital in Boston, Sylvain Bouix of the Superior Technology School in Canada, Richard Jarrett Rushmore of Boston University, Ebrahim Ebrahim of the technology company Kitware, and Professor Marc Niethammer, who advises Demir, Tian, and Greer. Tian is also a researcher at Massachusetts General Hospital.
The paper deals with image registration in medical imaging, which is the process of identifying the same part of the body in separate scans. Deep learning approaches have been developed for image registration, and these are typically faster but don’t generalize well to anatomies that weren’t part of the training data, and they are typically limited to one type of modality. Demir developed an approach to train a generalizable deep registration network that works within and across image modalities, for example allowing for registration between CT and CBCT images or between MR and CT, two use cases which are useful in radiation treatment planning.
More information about the project can be found on the project’s GitHub page.