Segmentation of scalp and skull in neonatal MR images using probabilistic atlas and level set method

Abstract

In this paper, we present a novel automatic algorithm for scalp and skull segmentation in T1-weighted neonatal head MR images. First, the probabilistic scalp and skull atlases are constructed. Second, the scalp outer surface is extracted based on an active mesh method. Third, maximum number of boundary points corresponding to the scalp inner surface is extracted using the constructed scalp probabilistic atlas and a set of knowledge based rules. In the next step, the skull inner surface and maximum number of boundary points of the outer surface are extracted using a priori information of the head anatomy and the constructed skull probabilistic atlas. Finally, the fast sweeping, tagging and level set methods are applied to reconstruct surfaces from the detected points in three-dimensional space. The results of the new segmentation algorithm on MRI data acquired from nine newborns (including three atlas and six test subjects) were compared with manual segmented data provided by an expert radiologist. The average similarity indices for the scalp and skull segmented regions were equal to 89% and 71% for the atlas and 84% and 63% for the test data, respectively.

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 .…”

“…Ghadimi et al [36] Similarity index [29] Similarity index Peporte et al [71] Dice metric, Jaccard metric, false positive, false negative Mahapatra [58] Dice metric, Jaccard index, sensitivity, specificity, false positive rate and Hausdorff distance 4.2.2.1. Atlas-based algorithms.…”

“…Ghadimi et al [36] used a two-step process for scalp and skull segmentation from T1-weighted MRI. First, the scalp and skull points were detected using a probabilistic atlas and constrained knowledge-based search.…”

“…Majority of the segmentation algorithms employ priori brain atlases or templates (Ghadimi, 2008 Wang, , 2014a. However, the performance of the atlas-based strategies is determined by the atlas quality and its correspondence with the subject (Shi, 2010a).…”

“…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.…”