Segmentation of Multiple Sclerosis plaques in magnetic resonance images using Gaussian mixture models and anatomical atlases

• Grant number: 12/16928-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): November 01, 2012
• Effective date (End): October 31, 2013
• Field of knowledge: Engineering - Biomedical Engineering - Bioengineering
• Principal Investigator: Ricardo José Ferrari
• Grantee: Bruno César Gregório da Silva
• Host Institution: Centro de Ciências Exatas e de Tecnologia (CCET). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil

Abstract

Multiple Sclerosis (MS) is an autoimmune, inflammatory, and demyelinating disease that attacks the central nervous system. It is considered a white matter (WM) disease because normally MS lesions appear in this area. Multispectral magnetic resonance imaging (MRI) has been used routinely to visually diagnose and monitor MS because of its excellent properties such as high resolution, good soft tissue differentiation, and different contrast information. However, visual assessment of MR images for the detection of MS lesions is a time-consuming and very tedious task. The use of probabilistic anatomical atlases has become a valuable tool in medical image processing. Such atlases provide spatial information of anatomical structures and thus enable the integration of this information in image processing algorithms. In this work, we propose to develop an automatic algorithm for the segmentation of MS lesions based on the Gaussian mixture model (GMM). Because MS lesions mainly appear in the white matter tissue, the probabilist anatomical atlases will be used to guide the estimation of the GMM as well as to constrain the WM region during image classification.(AU)