Automatic localization methodology dedicated to brain tumors based on ICP matching by using axial MRI symmetry

This research concerned one advanced methodology for automatic localization of brain tumors that could be imaged by Magnetic Resonance Image (MRI) modality. Such methodology could be based on Iterative closest point (ICP) matching technique by using axial MRI symmetry. The idea behind this work is to compare right and left hemispheres mirrored across a central axis. Indeed a healthy brain has a strong sagittal symmetry which is weakened by the presence of this distinguishable zone tumor. We propose a comparative study between this automatic methodology and the manual one for tumor localization. This could be done by comparing similarity between two sub-images representing respectively right and left hemisphere. The basic procedure was to extract the region of interest by using Maximum entropy thresholding and Otsu´s multilevel thresholding. Finally, we used the Iterative Closest Points (ICP) for the matching step.