Automatic corpus callosum segmentation using a deformable active Fourier contour model

Abstract

The corpus callosum (CC) is a structure of interest in many neuroimaging studies of neuro-developmental pathology such as autism. It plays an integral role in relaying sensory, motor and cognitive information from homologous regions in both hemispheres. We have developed a framework that allows automatic segmentation of the corpus callosum and its lobar subdivisions. Our approach employs constrained elastic deformation of flexible Fourier contour model, and is an extension of Szekely’s 2D Fourier descriptor based Active Shape Model. The shape and appearance model, derived from a large mixed population of 150+ subjects, is described with complex Fourier descriptors in a principal component shape space. Using MNI space aligned T1w MRI data, the CC segmentation is initialized on the mid-sagittal plane using the tissue segmentation. A multi-step optimization strategy, with two constrained steps and a final unconstrained step, is then applied. If needed, interactive segmentation can be performed via contour repulsion points. Lobar connectivity based parcellation of the corpus callosum can finally be computed via the use of a probabilistic CC subdivision model. Our analysis framework has been integrated in an open-source, end-to-end application called CCSeg both with a command line and Qt-based graphical user interface (available on NITRC). A study has been performed to quantify the reliability of the semi-automatic segmentation on a small pediatric dataset. Using 5 subjects randomly segmented 3 times by two experts, the intra-class correlation coefficient showed a superb reliability (0.99). CCSeg is currently applied to a large longitudinal pediatric study of brain development in autism.

Corpus Callosum Atrophy Rate in Mild Cognitive Impairment and Prodromal Alzheimer's Disease

“…1a) and has been shown to be highly correlated with CC volume, which is less easily measured because the lateral extents of the CC are not well-defined. Thus, the majority of CC studies in various conditions [7, 8, 14, 17, 20, 26] and nearly all proposed CC segmentation methods [27–31] target its mid-sagittal cross-sectional region.…”

Clustering High-Dimensional Landmark-Based Two-Dimensional Shape Data

“…Then, each T1-weighted MRI image and tissue segmentation calculated from FreeSurfer were used as the input files of CCSeg package to extract the planar CC shape data on the midsagittal slice, which contains 50 landmarks. The CCseg framework (Székely et al, 1996; Vachet et al, 2012) entails three main steps: (i) automatic initialization of the corpus callosum model, (ii) multi-step automatic (and potentially interactive) segmentation via constrained elastic deformation of a flexible Fourier contour model, and (iii) lobar area computation using a probabilistic subdivision model. After quality control, we obtained 647 CC shape data out of 776 subjects.…”

“…Subdivisions of corpus callosum in Witelson (1989) approach (left), its neuro-histological motivation (Vachet et al (2012), middle), and schematic visualization of the probability computation (Vachet et al (2012), right): prefrontal subdivision (blue); frontal subdivision (red); parietal (yellow); and occipito-temporal subdivision (green).…”

“…The majority of earlier work on shape analysis has focused on landmark-based analysis, where shapes are represented by a coarse, discrete sampling of the object contours (Bookstein, 1991; Cootes et al, 1995; Dryden and Mardia, 1998; Rajpoot and Arif, 2008). As an illustration, Figure 1 shows the automatic corpus callosum (CC) segmentations of four randomly selected subjects by using the CCSeg package 1 (Vachet et al, 2012). The CC contour (red) is represented by 100 landmarks, spacing along the contour about 0.75 mm.…”

A shape-template based two-stage corpus callosum segmentation technique for sagittal plane T1-weighted brain magnetic resonance images

“…Previous work on segmentation incorporating shape information in the Bspline framework include [13,14]. In our problem, CC has a distinguishable shape that makes it a suitable candidate for incorporating shape information into the segmentation algorithm [15,16]. On the energy front, many techniques rely on region-and gradient-based energies or functions derived thereof [17].…”