INSTITUTO DE FISICA LA PLATA
Unidad Ejecutora - UE
congresos y reuniones científicas
Diffuse Outlier Detection Technique for Functional Magnetic Resonance Imaging
JAVIER GIACOMANTONE; T. TARUTINA; ARMANDO DE GIUSTI
Congreso; XVI Congreso Argentino de Ciencias de la Computación. I Workshop: Procesamiento de Señales y Sistemas de Tiempo Real; 2010
We propose a new support vector machine (SVM) based method that improves the time series classication in magnetic resonance imaging (fMRI). We exploit the robust anisotropic diusion (RAD) technique to increase the classication performance of the one class support vector machine by taking into account the hypothesis of spatial relationshipbetween active voxels. The proposed method was called Diffuse OneClass Support Vector Machine (DOCSVM). DOCSVM method treats activated voxels as outliers and applies one class support vector machineto generate an activation map and RAD to include the neighborhoodhypothesis, improving the classication and reducing the iteration stepswith respect to RADSPM. We give a brief review of the main methods,present receiver operating characteristic (ROC) results and conclude suggesting further research alternatives.