Segmentación automática tridimensional de estructuras ventriculares cerebrales en imágenes de resonancia magnética/Three-dimensional automatic segmentation of ventricular brain structures in magnetic resonance imaging

A technique for brain ventricles structures 3-D segmentation is proposed using magnetic resonance images. This technique is global similarity enhancement-based technique and it comprises two stages: Filtering and Segmentation. During filtering a non-linear filters is used to address the noise y artifacts images problems. During segmentation, the region growing algorithm is applied to the pre-processed images and it is initialized using a voxel detected with least squares support vector machines. This kind of segmentation may be useful in the detection of several pathologies that impact the brain ventricles system.