Support Vector Machine (SVM) and the Neural Networks for Segmentation the Magnetic Resonance Imaging
Lahouaoui Lalaoui, Tayeb Mohamadi, Mohamed Chemachema, Abdessalem Hocini · 2009
In this paper a classification algorithms MLP and S VM for the segmentation tissue of brain magnetic resonance images is proposed. Magnetic resonance imaging (MRI) segmentation is an important technique to differentiate abnormal and normal tissues in MR ima ge data. The method interleaves classification with estimation of the model parameters, improving the classification at each iteration. This approach uses the informati on from the proton density (PD)- and T2-weighted and lattice relaxatio n time (T1) attenuation images. This works presents an NNN's and SVM (Support Vector Machine) for segmentation of mixed tissue in each voxels, for each voxels en calcu late the mean, variance and co-variance are the input of neural ne twork, to evaluate the performance on computed the misclassify rate image data and mixed tissue models. Segmentation tissue results from various algorithms are compared a nd the effectiveness and robustness of the proposed approa ch are demonstrated.