Segmentation of Magnetic Resonance Brain Tissues Image Based on Support Vector Machines and Level Set Method

MRI medical image segmentation is one of important problem in medical image processing. It is more challenging compared to other image processing problems due to the large variability in shapes, complexity of medical structures. In the paper, a new segmentation approach of magnetic resonance brain tissues image based on support vector machines (SVM) and level set method is presented. Firstly, reduced dimension based feature extraction followed by principal component analysis (PCA) is carried out and obtained results are used to train a SVM classifier. The result of classifier which closes to correct boundaries provides initial contours for the level set. Combined with SVM, level set method can achieve a refined and robust medical segmentation. The experimental results show that presented approach has faster convergence speed and better classification accuracy.