Implementation of a 3D method for computing texture parameters originated from the cooccurrence matrix in MR images

• Grant number: 15/11338-8
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: September 01, 2015
• End date: August 31, 2016
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Gabriela Castellano
• Grantee: João Pedro Oliveira Pompiani dos Santos
• Host Institution: Instituto de Física Gleb Wataghin (IFGW). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Associated research grant: 13/07559-3 - BRAINN - The Brazilian Institute of Neuroscience and Neurotechnology, AP.CEPID

Abstract

Texture analysis is an image processing technique that allows extraction of descriptors from an image, or a region of interest in the image, that are related to characteristics that refer to properties such as smoothness, rugosity, regularity etc. There are many ways to extract texture parameters of an image. For medical imaging, the most widely used approach has been the statistical approach, and within this, the class of parameters most frequently used has been the gray level cooccurrence matrix (MCO). Statistical parameters derived from the MCO have shown statistically significant differences in magnetic resonance imaging (MRI) of normal individuals and patients with different anomalies in brain structures that appear normal under a simple visual inspection of the image. Cooccurrence matrices are commonly computed for 2D slices of the MR image, which is 3D. The statistical parameters of each slice are then combined to give an estimate of the volumetric image's texture. This project aims to develop a method to calculate cooccurrence matrices directly from the 3D data, and, in this way, skip the step of combining the statistical parameters from the 2D slices, making the process more efficient and precise. Volumetrically, we can better characterize an alteration on a brain region of the patient, and, therefore, obtain texture parameters that can be more robust for comparison between different groups of people.