A Supervised Graph-Cut Deformable Model for Brain MRI Segmentation

Accurate automatic segmentation of subcortical brain structures in Magnetic Resonance Images (MRI) is of great interest in the analysis of developmental disorders. Segmentation methods based on a single or multiple atlases have been shown to suitably localize brain structures. However, the atlas prior information may not represent the structure of interest correctly. It may therefore be useful to introduce amore flexible technique for accurate segmentation. A fully-automatic segmentation method for brain MRI is considered, which defines a deformable model combining an atlas-based segmentation strategywith a supervisedGraph-cut model. TheGraphcut model is adapted to make it suitable for segmenting small and low-contrast brain structures by defining new data and boundary potentials of the energy function. L. Igual (B) · A. Hernández-Vela · S. Escalera · P. Radeva Universitat de Barcelona, Gran Via de les Corts Catalanes 585, 08007 Barcelona, Spain e-mail: [email protected] A. Hernández-Vela e-mail: [email protected] S. Escalera e-mail: [email protected] P. Radeva e-mail: [email protected] L. Igual · A. Hernández-Vela · S. Escalera · P. Radeva Computer Vision Center (CVC), Campus UAB, Edifici 0, Bellaterra, 08193 Barcelona, Spain J. C. Soliva · O. Vilarroya Unitat de Recerca en Neurociència Cognitiva (URNC) Department of Psychiatry, Universitat Autònoma de Barcelona (UAB), IAPS Hospital del Mar. Passeig Marítim, 25-29, 08003 Barcelona, Spain e-mail: [email protected] O. Vilarroya e-mail: [email protected] M. González Hidalgo et al. (eds.), Deformation Models, Lecture Notes in Computational 237 Vision and Biomechanics 7, DOI: 10.1007/978-94-007-5446-1_10, © Springer Science+Business Media Dordrecht 2013