Daniel N. Bullock

Daniel N. Bullock

Open science, data, A.I., and associated infrastructure policy, off-hours neuroinformatics researcher

Anatomically-Informed Multiple Linear Assignment Problems for White Matter Bundle Segmentation

Published in 2019 IEEE 16th International Symposium on Biomedical Imaging, 2019

Recommended citation: Bertò, G., Avesani, P., Pestilli, F., Bullock, D., Caron, B., & Olivetti, E. (2019, April). Anatomically-Informed Multiple Linear Assignment Problems for White Matter Bundle Segmentation. In 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019) (pp. 135-138). IEEE. https://arxiv.org/pdf/1907.07077

doi: https://doi.org/10.1007/978-3-030-04747-4_19

Segmenting white matter bundles from human tractograms is a task of interest for several applications. Current methods for bundle segmentation consider either only prior knowledge about the relative anatomical position of a bundle, or only its geometrical properties. Our aim is to improve the results of segmentation by proposing a method that takes into account information about both the underlying anatomy and the geometry of bundles at the same time. To achieve this goal, we extend a state-of-the-art example-based method based on the Linear Assignment Problem (LAP) by including prior anatomical information within the optimization process. The proposed method shows a significant improvement with respect to the original method, in particular on small bundles.