A hybrid geometric-statistical deformable model for automated 3-D segmentation in brain MRI

Whole brain segmentation with full volume neural network

Whole brain segmentation is an important neuroimaging task that segments the whole brain volume into anatomically labeled regions-of-interest. Convolutional neural networks have demonstrated good performance in this task. Existing solutions, usually segment the brain image by classifying the voxels, or labeling the slices or the sub-volumes separately. Their representation learning is based on parts of the whole volume whereas their labeling result is produced by aggregation of partial segmentation. Learning and inference with incomplete information could lead to sub-optimal final segmentation result. To address these issues, we propose to adopt a full volume framework, which feeds the full volume brain image into the segmentation network and directly outputs the segmentation result for the whole brain volume. The framework makes use of complete information in each volume and can be implemented easily. An effective instance in this framework is given subsequently. We adopt the 3D high-resolution network (HRNet) for learning spatially fine-grained representations and the mixed precision training scheme for memory-efficient training. Extensive experiment results on a publicly available 3D MRI brain dataset show that our proposed model advances the state-of-the-art methods in terms of segmentation performance.

A fast stochastic framework for automatic MR brain images segmentation

This paper introduces a new framework for the segmentation of different brain structures (white matter, gray matter, and cerebrospinal fluid) from 3D MR brain images at different life stages. The proposed segmentation framework is based on a shape prior built using a subset of co-aligned training images that is adapted during the segmentation process based on first- and second-order visual appearance characteristics of MR images. These characteristics are described using voxel-wise image intensities and their spatial interaction features. To more accurately model the empirical grey level distribution of the brain signals, we use a linear combination of discrete Gaussians (LCDG) model having positive and negative components. To accurately account for the large inhomogeneity in infant MRIs, a higher-order Markov-Gibbs Random Field (MGRF) spatial interaction model that integrates third- and fourth- order families with a traditional second-order model is proposed. The proposed approach was tested and evaluated on 102 3D MR brain scans using three metrics: the Dice coefficient, the 95-percentile modified Hausdorff distance, and the absolute brain volume difference. Experimental results show better segmentation of MR brain images compared to current open source segmentation tools.

A hyperspectral vessel image registration method for blood oxygenation mapping

Blood oxygenation mapping by the means of optical oximetry is of significant importance in clinical trials. This paper uses hyperspectral imaging technology to obtain in vivo images for blood oxygenation detection. The experiment involves dorsal skin fold window chamber preparation which was built on adult (8-10 weeks of age) female BALB/c nu/nu mice and in vivo image acquisition which was performed by hyperspectral imaging system. To get the accurate spatial and spectral information of targets, an automatic registration scheme is proposed. An adaptive feature detection method which combines the local threshold method and the level-set filter is presented to extract target vessels. A reliable feature matching algorithm with the correlative information inherent in hyperspectral images is used to kick out the outliers. Then, the registration images are used for blood oxygenation mapping. Registration evaluation results show that most of the false matches are removed and the smooth and concentrated spectra are obtained. This intensity invariant feature detection with outliers-removing feature matching proves to be effective in hyperspectral vessel image registration. Therefore, in vivo hyperspectral imaging system by the assistance of the proposed registration scheme provides a technique for blood oxygenation research.

Shape analysis of the human brain: a brief survey

The survey outlines and compares popular computational techniques for quantitative description of shapes of major structural parts of the human brain, including medial axis and skeletal analysis, geodesic distances, Procrustes analysis, deformable models, spherical harmonics, and deformation morphometry, as well as other less widely used techniques. Their advantages, drawbacks, and emerging trends, as well as results of applications, in particular, for computer-aided diagnostics, are discussed.

MRI segmentation of the human brain: challenges, methods, and applications

Image segmentation is one of the most important tasks in medical image analysis and is often the first and the most critical step in many clinical applications. In brain MRI analysis, image segmentation is commonly used for measuring and visualizing the brain's anatomical structures, for analyzing brain changes, for delineating pathological regions, and for surgical planning and image-guided interventions. In the last few decades, various segmentation techniques of different accuracy and degree of complexity have been developed and reported in the literature. In this paper we review the most popular methods commonly used for brain MRI segmentation. We highlight differences between them and discuss their capabilities, advantages, and limitations. To address the complexity and challenges of the brain MRI segmentation problem, we first introduce the basic concepts of image segmentation. Then, we explain different MRI preprocessing steps including image registration, bias field correction, and removal of nonbrain tissue. Finally, after reviewing different brain MRI segmentation methods, we discuss the validation problem in brain MRI segmentation.

Automated whole-brain N-acetylaspartate proton MRS quantification

Concentration of the neuronal marker, N-acetylaspartate (NAA), a quantitative metric for the health and density of neurons, is currently obtained by integration of the manually defined peak in whole-head proton ((1) H)-MRS. Our goal was to develop a full spectral modeling approach for the automatic estimation of the whole-brain NAA concentration (WBNAA) and to compare the performance of this approach with a manual frequency-range peak integration approach previously employed. MRI and whole-head (1) H-MRS from 18 healthy young adults were examined. Non-localized, whole-head (1) H-MRS obtained at 3 T yielded the NAA peak area through both manually defined frequency-range integration and the new, full spectral simulation. The NAA peak area was converted into an absolute amount with phantom replacement and normalized for brain volume (segmented from T1 -weighted MRI) to yield WBNAA. A paired-sample t test was used to compare the means of the WBNAA paradigms and a likelihood ratio test used to compare their coefficients of variation. While the between-subject WBNAA means were nearly identical (12.8 ± 2.5 mm for integration, 12.8 ± 1.4 mm for spectral modeling), the latter's standard deviation was significantly smaller (by ~50%, p = 0.026). The within-subject variability was 11.7% (±1.3 mm) for integration versus 7.0% (±0.8 mm) for spectral modeling, i.e., a 40% improvement. The (quantifiable) quality of the modeling approach was high, as reflected by Cramer-Rao lower bounds below 0.1% and vanishingly small (experimental - fitted) residuals. Modeling of the whole-head (1) H-MRS increases WBNAA quantification reliability by reducing its variability, its susceptibility to operator bias and baseline roll, and by providing quality-control feedback. Together, these enhance the usefulness of the technique for monitoring the diffuse progression and treatment response of neurological disorders.

A novel method of obtaining 3D images of detached retina

We describe a new method of obtaining three-dimensional (3D) images of detached retina. Twelve-slice photos of the partial retina were obtained according to the twelve positions on a clock face. Twelve sections were then cut from these photos and joined together. Each sector was resized to match nearby sectors and the complete retinal picture was then created. A sphere mapping algorithm was used to map the two-dimensional (2D) picture to a sphere, which was then used to simulate the actual eyeball. Finally, a 3D image of the entire retina with a clearly visible detached section was created by the Visualization Toolkit (VTK).

A learning-based wrapper method to correct systematic errors in automatic image segmentation: consistently improved performance in hippocampus, cortex and brain segmentation

We propose a simple but generally applicable approach to improving the accuracy of automatic image segmentation algorithms relative to manual segmentations. The approach is based on the hypothesis that a large fraction of the errors produced by automatic segmentation are systematic, i.e., occur consistently from subject to subject, and serves as a wrapper method around a given host segmentation method. The wrapper method attempts to learn the intensity, spatial and contextual patterns associated with systematic segmentation errors produced by the host method on training data for which manual segmentations are available. The method then attempts to correct such errors in segmentations produced by the host method on new images. One practical use of the proposed wrapper method is to adapt existing segmentation tools, without explicit modification, to imaging data and segmentation protocols that are different from those on which the tools were trained and tuned. An open-source implementation of the proposed wrapper method is provided, and can be applied to a wide range of image segmentation problems. The wrapper method is evaluated with four host brain MRI segmentation methods: hippocampus segmentation using FreeSurfer (Fischl et al., 2002); hippocampus segmentation using multi-atlas label fusion (Artaechevarria et al., 2009); brain extraction using BET (Smith, 2002); and brain tissue segmentation using FAST (Zhang et al., 2001). The wrapper method generates 72%, 14%, 29% and 21% fewer erroneously segmented voxels than the respective host segmentation methods. In the hippocampus segmentation experiment with multi-atlas label fusion as the host method, the average Dice overlap between reference segmentations and segmentations produced by the wrapper method is 0.908 for normal controls and 0.893 for patients with mild cognitive impairment. Average Dice overlaps of 0.964, 0.905 and 0.951 are obtained for brain extraction, white matter segmentation and gray matter segmentation, respectively.

The Alzheimer's disease neuroimaging initiative: progress report and future plans

The Alzheimer's Disease Neuroimaging Initiative (ADNI) beginning in October 2004, is a 6-year research project that studies changes of cognition, function, brain structure and function, and biomarkers in elderly controls, subjects with mild cognitive impairment, and subjects with Alzheimer's disease (AD). A major goal is to determine and validate MRI, PET images, and cerebrospinal fluid (CSF)/blood biomarkers as predictors and outcomes for use in clinical trials of AD treatments. Structural MRI, FDG PET, C-11 Pittsburgh compound B (PIB) PET, CSF measurements of amyloid beta (Abeta) and species of tau, with clinical/cognitive measurements were performed on elderly controls, subjects with mild cognitive impairment, and subjects with AD. Structural MRI shows high rates of brain atrophy, and has high statistical power for determining treatment effects. FDG PET, C-11 Pittsburgh compound B PET, and CSF measurements of Abeta and tau were significant predictors of cognitive decline and brain atrophy. All data are available at UCLA/LONI/ADNI, without embargo. ADNI-like projects started in Australia, Europe, Japan, and Korea. ADNI provides significant new information concerning the progression of AD.