We propose a method for the segmentation of Multiple Sclerosis lesions. The method is based on probability maps derived from a K-Nearest Neighbours classication. These are used as a non parametric likelihood in a Bayesian formulation with a prior that assumes connectivity of neighbouring voxels. The formulation is solved using the method of Iterated Conditional Modes (ICM). The parameters of the method are found through leave-one-out cross validation on training data after which it is evaluated on previously unseen test data. The multi modal features investigated are 3 structural MRI modalities, the diusion MRI measures of Fractional Anisotropy (FA), Mean Diusivity (MD) and several spatial features. Results show a benet from the inclusion of diusion primarily to the most dicult cases. Results shows that combining probabilistic K-Nearest Neighbour with a Markov Random Field formulation leads to a slight improvement of segmentations.
|Title of host publication||ICIAR Proceedings : Springer Lecture Notes|
|Number of pages||8|
|Publication status||Published - 2012|
|Event||International Conference on Image Analysis and Recognition, ICIAR 2012 - Aveiro, Portugal|
Duration: 25 Jun 2012 → 27 Jun 2012
|Conference||International Conference on Image Analysis and Recognition, ICIAR 2012|
|Period||25/06/2012 → 27/06/2012|
Lyksborg, M., Larsen, R., Sørensen, P. S., Blinkenberg, M., Garde, E., Siebner, H. R., & Dyrby, T. B. (2012). Segmenting Multiple Sclerosis Lesions using a Spatially Constrained K-Nearest Neighbour approach. In ICIAR Proceedings: Springer Lecture Notes Springer.