An approach to classifying Magnetic Resonance (MR) image data is described. The specific application is the classification of MRI scan data according to the nature of the corpus callosum, however the approach has more general applicability. A variation of the "spectral segmentation with multi-scale graph decomposition" mechanism is introduced. The result of the segmentation is stored in a quad-tree data structure to which a weighted variation (also developed by the authors) of the gSpan algorithm is applied to identify frequent sub-trees. As a result the images are expressed as a set frequent sub-trees. There may be a great many of these and thus a decision tree based feature reduction technique is applied before classification takes place. The results show that the proposed approach performs both efficiently and effectively, obtaining a classification accuracy of over 95% in the case of the given application. © 2010 Springer-Verlag London.
CITATION STYLE
Elsayed, A., Coenen, F., Jiang, C., García-Fiñana, M., & Sluming, V. (2010). Corpus callosum MR image classification. In Research and Development in Intelligent Systems XXVI: Incorporating Applications and Innovations in Intelligent Systems XVII (pp. 333–346). Springer London. https://doi.org/10.1007/978-1-84882-983-1_27
Mendeley helps you to discover research relevant for your work.