Classifier system for normality pre-diagnosis from MRI brain images in cerebral pathology studies

Resumen

This research work presents the modeling and development of a Classifier System applied as a support tool to pre-diagnoses of normality/abnormality in brain pathology studies from TC or MRI digital images of a human brain. The proposed model is composed of three phases (1) images pre-processing, (2) features extraction basis on Gabor Filtering and (3) a classifier using Support Vector Machine with various kernel. This system has been tested on real MRI brain images obtaining a classification rate of 95% using RBF kernel.

Christian E. Portugal-Zambrano

Director de Tecnologías en Jebi S.A.C

Mis intereses de investigación incluyen la Inteligencia Artificial, procesamiento de imágenes e internet de las cosas.