An extension Gaussian mixture model for brain MRI segmentation

The segmentation of brain magnetic resonance (MR) images into gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) has been an intensive studied area in the medical image analysis community. The Gaussian mixture model (GMM) is one of the most commonly used model to represent the intensity of different tissue types. However, as a histogram-based model, the spatial relationship between pixels is discarded in the GMM, making it sensitive to noise. Herein we present a new framework which aims to incorporate spatial information into the standard GMM, where each pixel is assigned its individual prior by leveraging its neighborhood information. Expectation maximization (EM) is modified to estimate the parameters of the proposed method. The method is validated on both synthetic and real brain MR images, showing its effectiveness in the segmentation task.

[1]  Donald Geman,et al.  Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images , 1984 .

[2]  Ron Kikinis,et al.  Markov random field segmentation of brain MR images , 1997, IEEE Transactions on Medical Imaging.

[3]  Ben J. A. Kröse,et al.  Efficient Greedy Learning of Gaussian Mixture Models , 2003, Neural Computation.

[4]  Abdul Rahman Ramli,et al.  Review of brain MRI image segmentation methods , 2010, Artificial Intelligence Review.

[5]  Theo Gevers,et al.  A Spatially Constrained Generative Model and an EM Algorithm for Image Segmentation , 2007, IEEE Transactions on Neural Networks.

[6]  Reza Azmi,et al.  Brain tissue segmentation in MR images based on a hybrid of MRF and social algorithms , 2012, Medical Image Anal..

[7]  Chunming Li,et al.  MRI Tissue Classification and Bias Field Estimation Based on Coherent Local Intensity Clustering: A Unified Energy Minimization Framework , 2009, IPMI.

[8]  Hui Zhang,et al.  Image segmentation by a robust Modified Gaussian Mixture Model , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[9]  M. Mafee,et al.  Role of MR Imaging , 1999 .

[10]  Koenraad Van Leemput,et al.  Automated model-based bias field correction of MR images of the brain , 1999, IEEE Transactions on Medical Imaging.

[11]  Bostjan Likar,et al.  A Review of Methods for Correction of Intensity Inhomogeneity in MRI , 2007, IEEE Transactions on Medical Imaging.

[12]  Stephen M. Smith,et al.  Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm , 2001, IEEE Transactions on Medical Imaging.

[13]  Q. M. Jonathan Wu,et al.  An Extension of the Standard Mixture Model for Image Segmentation , 2010, IEEE Transactions on Neural Networks.

[14]  G. McLachlan,et al.  The EM algorithm and extensions , 1996 .

[15]  Qiang Chen,et al.  Fuzzy c-means clustering with weighted image patch for image segmentation , 2012, Appl. Soft Comput..

[16]  Q. M. Jonathan Wu,et al.  Fast and Robust Spatially Constrained Gaussian Mixture Model for Image Segmentation , 2013, IEEE Transactions on Circuits and Systems for Video Technology.

[17]  J. Mazziotta,et al.  Brain Mapping: The Methods , 2002 .

[18]  Y. Ge.  Multiple sclerosis: the role of MR imaging. , 2006, AJNR. American journal of neuroradiology.

[19]  Qiang Chen,et al.  Fuzzy Local Gaussian Mixture Model for Brain MR Image Segmentation , 2012, IEEE Transactions on Information Technology in Biomedicine.