Development of novel framework for brain tumor segmentation on MRI
Development of a novel validation framework for brain tumor segmentation. Required in neurosurgery planning, radio- and chemotherapy monitoring, drug discovery…
Approach
- New brain segmentation algorithm derived from spectral clustering algorithms. Segmentation as weighted graph partitioning. Previous work of Shi and Malik 2000, NCUT criterion, Ng, Jordan and Weiss 2002, Supervised clustering using k eigenvectors, Miela and Shi 2002, Supervised clustering – connection with Markov Chains Fowlkes, Belongie, Chung, Malik 2004, Nyström method – spine segmentation from MRI, Fiedler eigenvector based segmentation, Archip et al. 2005.
- Estimate the performance of segmentation algorithms (STAPLE)
