Skull segmentation in 3D neonatal MRI using hybrid Hopfield Neural Network

Abstract

A fully automated method for segmentation of neonatal skull in Magnetic Resonance (MR) images for source localization of electrical/magnetic encephalography (EEG/MEG) signals is proposed. Finding the source of these signals shows the origin of an abnormality. We propose a hybrid algorithm in which a Bayesian classifying framework is combined with a Hopfield Neural Network (HNN) for neonatal skull segmentation. Due to the non-homogeneity of skull intensities in MR images, local statistical parameters are used for adaptive training of Hopfield neural network based on Bayesian classifier error. The experimental results, which are obtained on high resolution T1-weighted MR images of nine neonates with gestational ages between 39 and 42 weeks, show 65% accuracy which consistently exhibits our scheme's superiority in comparison with previous neonatal skull segmentation methods.

Neonatal brain MRI segmentation: A review

“…Nevertheless most algorithms have reported good results on par or better than previously published works. In particular, Daliri et al [29] have shown an improvement in SI values compared to Ghadimi et al [36]; Peporte et al [71] have also reported better performance of their approach compared to four publicly available brain extraction tools, namely, BrainSuite 6 , SPM8-Statistical Parametric Mapping 7 , FMRIB Software Library (FSL) 8 and MRIcroN 9 . Likewise, Mahapatra [58] has demonstrated better accuracy of their results in relation to the graph cut approach of Sadananthan et al [83] and other publicly available software like BET [97], BSE [89] and Hybrid Watershed Algorithm (HWA) [86] that is a part of FreeSurfer Software Suite 10 .…”

“…This was followed by a level set based reconstruction to obtain closed surfaces. One other work that employed a probabilistic atlas for skull segmentation is that of Daliri et al [29]. This approach proposed an adaptive algorithm using Bayesian classifier to weight local features of the MR image against those of the atlas.…”

“…portion is the first step in any brain segmentation algorithm (Makropoulos, 2012; Melbourne, 2012; Egekher, 2012), some algorithms focus exclusively on skull stripping (Daliri, 2010;[58]). Such approaches are limited in the sense that they cannot segment brain tissues such as the GM, WM and CSF.…”

“…While a few algorithms merely skull strip the brain images (Ghadimi, 2008;Daliri, 2010; [58]), most of them are capable of identifying the prominent brain portions such as GM, WM and cerebrospinal fluid (CSF) [100]; some are adept at further segmenting the brain into eight tissue classes (Egekher, 2012; Anbeek, 2013) while a recent research has been successful in parcellating the brain into as many as 50 regions (Gousias, 2013). Given the vast difference in the degree of tissue segmentation, the algorithms also vary in level and complexity.…”

A Novel FCM Algorithm Incorporating Spatial Information for Color Image Segmentation

“…Image segmentation plays an important role in variety of applications such as robot vision, object recognition, geographical imaging and medical imaging [1][2][3][4][5][6]. In general, fuzzy segmentation methods, especially the FCM algorithm, have been widely used in image segmentation [7][8][9].…”