Segmentation of subcortical brain structures using fuzzy templates

FreeSurfer-initiated fully-automated subcortical brain segmentation in MRI using Large Deformation Diffeomorphic Metric Mapping

“…Looking at recently published segmentation methods, FS+LDDMM demonstrates comparable levels of accuracy as indicated by spatial overlap: Pitiot et al (2004) reported an average mean surface distance of 1.6 mm for the caudate nucleus; our mean surface distances ranged from 0.65 mm to 1.01 mm. Zhou and Rajapakse (2005) reported spatial overlaps (DSC) of 0.81, 0.84 and 0.83 for the caudate, putamen, and thalamus respectively, and Amini et al (2004) reported a spatial overlap (DSC) of 0.88 for the thalamus, figures which are very close in magnitude with our results (DSC = 0.81, 0.83 and 0.86 for the caudate, putamen, and thalamus in healthy controls). Methods which are tailored for the segmentation of a specific structure (Xia et al, 2007;Chupin et al, 2007) are likely to achieve higher spatial overlap, such as those presented by Xia et al (2007) (DSC = 0.873 ± 0.0234 for the caudate nucleus), however, this knowledge-driven method may be unlikely to achieve the same results on other datasets or pathologies and is usable only on the caudate nucleus.…”

Patch-based segmentation using expert priors: Application to hippocampus and ventricle segmentation

“…In contrast to brain tissue classification where the intensity of the MR signal can be used to segment different tissue types, anatomical segmentation usually requires information derived from the manual segmentations done by experts (i.e., expert priors), since anatomical structures can be composed of several tissue types and distinct anatomical structures can have the same MR signal properties. To overcome this difficulty, several automatic methods of segmentation have been proposed, such as deformable models or region growing (Chupin et al, 2007;Ghanei et al, 1998;Shen et al, 2002), appearance-based models (Duchesne et al, 2002;Hu and Collins, 2007), and atlas/template-warping techniques (Aljabar et al, 2009;Barnes et al, 2008;Collins et al, 1995;Fischl et al, 2002;Gousias et al, 2008;Hammers et al, 2007;Heckemann et al, 2006;Rohlfing et al, 2004;Zhou and Rajapakse, 2005).…”

Markov Dependence Tree-Based Segmentation of Deep Brain Structures

“…We focus on the segmentation of six gray matter brain structures, i.e., left and right caudate nuclei, putamina and thalami. These six deep brain structures are often studied in brain anatomy [2,3,4,5,8,10,11,12,13,15]. However, they are difficult to segment automatically due to their blurry boundary and small sizes.…”

“…Due to its linearity, PCA may not be able to describe the relative positions of multiple objects and their non-linear variations. The second popular brain structure segmentation strategy is fuzzy logic control [6,7,8], which can manage the selection of various candidates of possible structures. The problem with fuzzy logic is that it is difficult to give consistently high accuracy to the segmentations of various intracranial structures because the relationship among those structures maintained by fuzzy logic may be weak and imprecise.…”

“…One common feature of the above-mentioned methods [3,4,5,7,8,9,10,11,12,13,14] is the need of sufficient training samples. The training process usually requires a considerable number of segmented images with the known ground truths.…”