A Graph Based Classification Method for Multiple Sclerosis Clinical Forms Using Support Vector Machine

摘要

With the development of advanced image acquisition and processing techniques providing better biomarkers for the characterization of brain diseases, automatic classification of biomedical imaging becomes an important field in research. Since brain neural network is one of the most complex network, graph theory constitutes a promising approach to characterize its connectivity properties. In this work, we applied this technique to diffusion tensor imaging data acquired in multiple sclerosis (MS) patients in order to classify their clinical forms. Support Vector Machine (SVM) algorithm in combination with graph kernel were used to classify 65 MS patients in three different clinical forms. Results showed high classification performances using both weighted and unweighted connectivity graphs, the later being more stable, and less dependent to the pathological conditions.