Optimal Statistical Structure Validation of Brain Tumors Using Refractive Index

Tumour segmentation from brain MRI is more than a decade old problem in the field of medical imaging. Till date automated brain tumour segmentation happens to be a difficult task due to the variance and complexity of tumour growth. In this paper, we present this segmentation problem for the purpose of determining the exact location of brain tumour using refractive index study on the structural analysis of both tumorous and normal tissues. Initially, as per existing survey 3 kinds of features namely, intensity-based, texture-based, and symmetry-based are extracted from the structural elements. Then reduction of this feature set is performed and similar features are clustered together. Refractive index analysis is performed on each of the clusters from the MR T2 relaxation time. Deviation from a threshold value of RI for majority of pixels in a particular cluster denotes it to be the tumorous region.