Segmentation and interpretation of MR brain images: an improved active shape model

Mindboggle: a scatterbrained approach to automate brain labeling

Mindboggle (http://www.binarybottle.com/mindboggle.html) is a fully automated, feature matching approach to label cortical structures and activity anatomically in human brain MRI data. This approach does not assume that the existence of component structures and their relative spatial relationship is preserved from brain to brain, but instead disassembles a labeled atlas and reassembles its pieces to match corresponding pieces in an unlabeled subject brain before labeling. Mindboggle: (1) converts linearly coregistered subject and atlas MRI data into sulcus pieces, (2) matches each atlas piece with a combination of subject pieces by minimizing a cost function, (3) transforms atlas label boundaries to the matching subject pieces, (4) warps atlas labels to their transformed boundaries, and (5) propagates labels to fill remaining gaps in a mask derived from the subject brain. We compared Mindboggle with four registration methods: linear registration, and nonlinear registration using SPM2, AIR, and ANIMAL. Automated labeling by all of the nonlinear methods was found to be at least comparable with linear registration. Mindboggle outperformed every other method, as measured by the agreement between overlapping atlas labels and manually assigned subject labels, with respect to the union or the intersection of voxels. After applying the same procedure that Mindboggle uses to fill a subject's segmented gray matter mask with labels (step 5), the results of the other methods improved. However, after performing a one-way ANOVA (and Tukey's honestly significant difference criterion) in a multiple comparison between the results obtained by the different methods, Mindboggle was still found to be the only nonlinear method whose labeling performance was significantly better than that of linear registration or SPM2. Further advantages to Mindboggle include a high degree of robustness against image artifacts, poor image quality, and incomplete brain data. We tested the latter hypothesis by conducting all of the tests again, this time registering the atlas to an artificially lesioned version of itself, and found that Mindboggle was the only method whose performance did not degrade significantly as the lesion size increased.

Determining correspondence in 3-D MR brain images using attribute vectors as morphological signatures of voxels

Finding point correspondence in anatomical images is a key step in shape analysis and deformable registration. This paper proposes an automatic correspondence detection algorithm for intramodality MR brain images of different subjects using wavelet-based attribute vectors (WAVs) defined on every image voxel. The attribute vector (AV) is extracted from the wavelet subimages and reflects the image structure in a large neighborhood around the respective voxel in a multiscale fashion. It plays the role of a morphological signature for each voxel, and our goal is, therefore, to make it distinctive of the respective voxel. Correspondence is then determined from similarities of AVs. By incorporating the prior knowledge of the spatial relationship among voxels, the ability of the proposed algorithm to find anatomical correspondence is further improved. Experiments with MR images of human brains show that the algorithm performs similarly to experts, even for complex cortical structures.

Classification of Anatomical Structures in MR Brain Images Using Fuzzy Parameters

We present an algorithm that automatically segments and classifies the brain structures in a set of magnetic resonance (MR) brain images using expert information contained in a small subset of the image set. The algorithm is intended to do the segmentation and classification tasks mimicking the way a human expert would reason. The algorithm uses a knowledge base taken from a small subset of semiautomatically classified images that is combined with a set of fuzzy indexes that capture the experience and expectation a human expert uses during recognition tasks. The fuzzy indexes are tissue specific and spatial specific, in order to consider the biological variations in the tissues and the acquisition inhomogeneities through the image set. The brain structures are segmented and classified one at a time. For each brain structure the algorithm needs one semiautomatically classified image and makes one pass through the image set. The algorithm uses low-level image processing techniques on a pixel basis for the segmentations, then validates or corrects the segmentations, and makes the final classification decision using higher level criteria measured by the set of fuzzy indexes. We use single-echo MR images because of their high volumetric resolution; but even though we are working with only one image per brain slice, we have multiple sources of information on each pixel: absolute and relative positions in the image, gray level value, statistics of the pixel and its three-dimensional neighborhood and relation to its counterpart pixels in adjacent images. We have validated our algorithm for ease of use and precision both with clinical experts and with measurable error indexes over a Brainweb simulated MR set.

Automatic segmentation of thalamus from brain MRI integrating fuzzy clustering and dynamic contours

Thalamus is an important neuro-anatomic structure in the brain. In this paper, an automated method is presented to segment thalamus from magnetic resonance images (MRI). The method is based on a discrete dynamic contour model that consists of vertices and edges connecting adjacent vertices. The model starts from an initial contour and deforms by external and internal forces. Internal forces are calculated from local geometry of the model and external forces are estimated from desired image features such as edges. However, thalamus has low contrast and discontinues edges on MRI, making external force estimation a challenge. The problem is solved using a new algorithm based on fuzzy C-means (FCM) unsupervised clustering, Prewitt edge-finding filter, and morphological operators. In addition, manual definition of the initial contour for the model makes the final segmentation operator-dependent. To eliminate this dependency, new methods are developed for generating the initial contour automatically. The proposed approaches are evaluated and validated by comparing automatic and radiologist's segmentation results and illustrating their agreement.

Interactive segmentation of abdominal aortic aneurysms in CTA images

A model-based approach to interactive segmentation of abdominal aortic aneurysms from CTA data is presented. After manual delineation of the aneurysm sac in the first slice, the method automatically detects the contour in subsequent slices, using the result from the previous slice as a reference. If an obtained contour is not sufficiently accurate, the user can intervene and provide an additional manual reference contour. The method is inspired by the active shape model (ASM) segmentation scheme (), in which a statistical shape model, derived from corresponding landmark points in manually labeled training images, is fitted to the image in an iterative manner. In our method, a shape model of the contours in two adjacent image slices is progressively fitted to the entire volume. The contour obtained in one slice thus constrains the possible shapes in the next slice. The optimal fit is determined on the basis of multi-resolution gray level models constructed from gray value patches sampled around each landmark. We propose to use the similarity of adjacent image slices for this gray level model, and compare these to single-slice features that are more generally used with ASM. The performance of various image features is evaluated in leave-one-out experiments on 23 data sets. Features that use the similarity of adjacent image slices outperform measures based on single-slice features in all cases. The average number of slices in our datasets is 51, while on average eight manual initializations are required, which decreases operator segmentation time by a factor of 6.

Automated human frontal lobe identification in MR images based on fuzzy-logic encoded expert anatomic knowledge

Identification of human brain structures in MR images comprises an area of increasing interest, which also presents numerous methodological challenges. Here we describe a new knowledge-based automated method designed to identify several major brain sulci and then to define the frontal lobes by using the identified sulci as landmarks. To identify brain sulci, sulcal images were generated by morphologic operations and then separated into different components based on connectivity analysis. Subsequently, the individual anatomic features were evaluated by using fuzzy membership functions. The crisp decisions, i.e., the identification of sulci, were made by taking the maximum of the summation of all the membership functions. The identification was designed in a hierarchical order. The longitudinal fissure was extracted first. The left and right central sulci were then identified based on the left and right hemispheres. Next, the lateral sulci were identified based on the central sulci and hemispheres. Finally, the left and right frontal lobes were defined from the two hemispheres. The method was evaluated by visual inspection, comparison with manual segmentation, and comparison with manually volumetric results in references. The average Jaccard similarities of left and right frontal lobes between the automated and manual segmentation were 0.89 and 0.91, respectively. The average Kappa indices of left and right frontal lobes between the automated and manual segmentation were 0.94 and 0.95, respectively. These results show relatively high accuracy of using this novel method for human frontal lobe identification and segmentation.

Segmentation of magnetic resonance brain images through discriminant analysis

Segmentation (tissue classification) of medical images obtained from a magnetic resonance (MR) system is a primary step in most applications of medical image post-processing. This paper describes nonparametric discriminant analysis methods to segment multispectral MR images of the brain. Starting from routinely available spin-lattice relaxation time, spin-spin relaxation time, and proton density weighted images (T1w, T2w, PDw), the proposed family of statistical methods is based on: (i) a transform of the images into components that are statistically independent from each other; (ii) a nonparametric estimate of probability density functions of each tissue starting from a training set; (iii) a classic Bayes 0-1 classification rule. Experiments based on a computer built brain phantom (brainweb) and on eight real patient data sets are shown. A comparison with parametric discriminant analysis is also reported. The capability of nonparametric discriminant analysis in improving brain tissue classification of parametric methods is demonstrated. Finally, an assessment of the role of multispectrality in classifying brain tissues is discussed.

Structural shape characterization via exploratory factor analysis

This article presents an exploratory factor analytic approach to morphometry in which a high-dimensional set of shape-related variables is examined with the purpose of finding clusters with strong correlation. This clustering can potentially identify regions that have anatomic significance and thus lend insight to knowledge discovery and morphometric investigations.

METHODS

The information about regional shape is extracted by registering a reference image to a set of test images. Based on the displacement fields obtained form image registration, the amount of pointwise volume enlargement or reduction is computed and statistically analyzed with the purpose of extracting a reduced set of common factors.

EXPERIMENTS

The effectiveness and robustness of the method is demonstrated in a study of gender-related differences of the human corpus callosum anatomy, based on a sample of 84 right-handed normal controls.

RESULTS

The method is able to automatically partition the structure into regions of interest, in which the most relevant shape differences can be observed. The confidence of results is evaluated by analyzing the statistical fit of the model and compared to previous experimental works.

Tracking the optic nervehead in OCT video using dual eigenspaces and an adaptive vascular distribution model

Optical coherence tomography (OCT) is a new ophthalmic imaging modality generating cross sectional views of the retina. OCT systems are essentially Michelson interferometers that form images in 1.5 s by directing a superluminescent diode (SLD) beam over the retinal surface. Involuntary eye motions frequently cause incorrect locations to be imaged. This motion may leave no obvious artifacts in the scan data and can easily go undetected. For glaucoma monitoring especially, knowing the measurement path, typically a circle concentric with the nerve head, is crucial. The commercially available OCT system displays a near-infrared video of the retina showing the SLD beam. This paper presents a prototype system to detect the nerve head and SLD beam in the video, and report the true scan path relative to the nerve head. Low image contrast and limited resolution make the reliable detection of retinal features difficult. In an adaptive model construction phase, the system directly detects retinal vasculature and the nerve head and incrementally builds a model of the current subject's vascular pattern relative to the optic disk. The nerve head identification is multitiered, using a novel dual eigenspace technique and a geometric comparison of detected vessel positions and nerve head hypotheses. In its operational phase, a correspondence is achieved between the currently detected vasculature and the model. Using subjects not included in training, the system located the optic nerve head to within 5 pixels (0.07 optic disk diameters, an error well below clinical significance) in 99.75% of 2800 video fields. In current clinical practice, motions as large as 1-2 disc diameters may go undetected, so this is a vast improvement.

Computer-assisted enhanced volumetric segmentation magnetic resonance imaging data using a mixture of artificial neural networks

An accurate computer-assisted method able to perform regional segmentation on 3D single modality images and measure its volume is designed using a mixture of unsupervised and supervised artificial neural networks. Firstly, an unsupervised artificial neural network is used to estimate representative textures that appear in the images. The region of interest of the resultant images is selected by means of a multi-layer perceptron after a training using a single sample slice, which contains a central portion of the 3D region of interest. The method was applied to magnetic resonance imaging data collected from an experimental acute inflammatory model (T(2) weighted) and from a clinical study of human Alzheimer's disease (T(1) weighted) to evaluate the proposed method. In the first case, a high correlation and parallelism was registered between the volumetric measurements, of the injured and healthy tissue, by the proposed method with respect to the manual measurements (r = 0.82 and p < 0.05) and to the histopathological studies (r = 0.87 and p < 0.05). The method was also applied to the clinical studies, and similar results were derived of the manual and semi-automatic volumetric measurement of both hippocampus and the corpus callosum (0.95 and 0.88).

Adapting Active Shape Models for 3D segmentation of tubular structures in medical images

Active Shape Models (ASM) have proven to be an effective approach for image segmentation. In some applications, however, the linear model of gray level appearance around a contour that is used in ASM is not sufficient for accurate boundary localization. Furthermore, the statistical shape model may be too restricted if the training set is limited. This paper describes modifications to both the shape and the appearance model of the original ASM formulation. Shape model flexibility is increased, for tubular objects, by modeling the axis deformation independent of the cross-sectional deformation, and by adding supplementary cylindrical deformation modes. Furthermore, a novel appearance modeling scheme that effectively deals with a highly varying background is developed. In contrast with the conventional ASM approach, the new appearance model is trained on both boundary and non-boundary points, and the probability that a given point belongs to the boundary is estimated non-parametrically. The methods are evaluated on the complex task of segmenting thrombus in abdominal aortic aneurysms (AAA). Shape approximation errors were successfully reduced using the two shape model extensions. Segmentation using the new appearance model significantly outperformed the original ASM scheme; average volume errors are 5.1% and 45% respectively.

Optimal deformable surface models for 3D medical image analysis

We present a novel method for building an optimal statistical deformable model from a set of surfaces whose topological realization is homeomorphic to a compact 2D manifold with boundary. The optimal parameterization of each shape is recursively refined by using hierarchical PBMs and tensor product B-spline representation of the surface. A criterion based on MDL is used to define the internal correspondence of the training data. The strength of the proposed technique is demonstrated by deriving a concise statistical model of the human left ventricle which has principal modes of variation that correspond to intrinsic cardiac motions. We demonstrate how the derived model can be used for 3D dynamic volume segmentation of the left ventricle, with its accuracy assessed by comparing results obtained from manual delineation of 3D cine MR data of 8 asymptomatic subjects. The extension of the technique to shapes with complex topology is also discussed.

Segmentation of prostate boundaries from ultrasound images using statistical shape model

This paper presents a statistical shape model for the automatic prostate segmentation in transrectal ultrasound images. A Gabor filter bank is first used to characterize the prostate boundaries in ultrasound images in both multiple scales and multiple orientations. The Gabor features are further reconstructed to be invariant to the rotation of the ultrasound probe and incorporated in the prostate model as image attributes for guiding the deformable segmentation. A hierarchical deformation strategy is then employed, in which the model adaptively focuses on the similarity of different Gabor features at different deformation stages using a multiresolution technique, i.e., coarse features first and fine features later. A number of successful experiments validate the algorithm.

Combining low-, high-level and empirical domain knowledge for automated segmentation of ultrasonic breast lesions

Breast cancer is the most frequently diagnosed malignancy and the second leading cause of mortality in women. In the last decade, ultrasound along with digital mammography has come to be regarded as the gold standard for breast cancer diagnosis. Automatically detecting tumors and extracting lesion boundaries in ultrasound images is difficult due to their specular nature and the variance in shape and appearance of sonographic lesions. Past work on automated ultrasonic breast lesion segmentation has not addressed important issues such as shadowing artifacts or dealing with similar tumor like structures in the sonogram. Algorithms that claim to automatically classify ultrasonic breast lesions, rely on manual delineation of the tumor boundaries. In this paper, we present a novel technique to automatically find lesion margins in ultrasound images, by combining intensity and texture with empirical domain specific knowledge along with directional gradient and a deformable shape-based model. The images are first filtered to remove speckle noise and then contrast enhanced to emphasize the tumor regions. For the first time, a mathematical formulation of the empirical rules used by radiologists in detecting ultrasonic breast lesions, popularly known as the "Stavros Criteria" is presented in this paper. We have applied this formulation to automatically determine a seed point within the image. Probabilistic classification of image pixels based on intensity and texture is followed by region growing using the automatically determined seed point to obtain an initial segmentation of the lesion. Boundary points are found on the directional gradient of the image. Outliers are removed by a process of recursive refinement. These boundary points are then supplied as an initial estimate to a deformable model. Incorporating empirical domain specific knowledge along with low and high-level knowledge makes it possible to avoid shadowing artifacts and lowers the chance of confusing similar tumor like structures for the lesion. The system was validated on a database of breast sonograms for 42 patients. The average mean boundary error between manual and automated segmentation was 6.6 pixels and the normalized true positive area overlap was 75.1%. The algorithm was found to be robust to 1) variations in system parameters, 2) number of training samples used, and 3) the position of the seed point within the tumor. Running time for segmenting a single sonogram was 18 s on a 1.8-GHz Pentium machine.

A new deformable model for analysis of X-ray CT images in preclinical studies of mice for polycystic kidney disease

This paper describes the application of a new probabilistic shape and appearance model (PSAM) algorithm to the task of detecting polycystic kidney disease (PKD) in X-ray computed tomography images of laboratory mice. The genetically engineered PKD mouse is a valuable animal model that can be used to develop new treatments for kidney-related problems in humans. PSAM is a statistical-based deformable model that improves upon existing point distribution models for boundary-based object segmentation. This new deformable model algorithm finds the optimal boundary position using an objective function that has several unique characteristics. Most importantly, the objective function includes both global shape and local gray-level characteristics, so optimization occurs with respect to both pieces of information simultaneously. PSAM is employed to segment the mouse kidneys and then texture measurements are applied within kidney boundaries to detect PKD. The challenges associated with the segmentation non-rigid organs along with the availability of a priori information led to the choice of a trainable, deformable model for this application. In 103 kidney images that were analyzed as part of a preclinical animal study, the mouse kidneys and spine were segmented with an average error of 2.4 pixels per boundary point. In all 103 cases, the kidneys were successfully segmented at a level where PKD could be detected using mean-of-local-variance texture measurements within the located boundary.

3-D active appearance models: segmentation of cardiac MR and ultrasound images

A model-based method for three-dimensional image segmentation was developed and its performance assessed in segmentation of volumetric cardiac magnetic resonance (MR) images and echocardiographic temporal image sequences. Comprehensive design of a three-dimensional (3-D) active appearance model (AAM) is reported for the first time as an involved extension of the AAM framework introduced by Cootes et al. The model's behavior is learned from manually traced segmentation examples during an automated training stage. Information about shape and image appearance of the cardiac structures is contained in a single model. This ensures a spatially and/or temporally consistent segmentation of three-dimensional cardiac images. The clinical potential of the 3-D AAM is demonstrated in short-axis cardiac MR images and four-chamber echocardiographic sequences. The method's performance was assessed by comparison with manually identified independent standards in 56 clinical MR and 64 clinical echo image sequences. The AAM method showed good agreement with the independent standard using quantitative indexes of border positioning errors, endo- and epicardial volumes, and left ventricular mass. In MR, the endocardial volumes, epicardial volumes, and left ventricular wall mass correlation coefficients between manual and AAM were R2 = 0.94, 0.97, 0.82, respectively. For echocardiographic analysis, the area correlation was R2 = 0.79. The AAM method shows high promise for successful application to MR and echocardiographic image analysis in a clinical setting.

Active shape model segmentation with optimal features

An active shape model segmentation scheme is presented that is steered by optimal local features, contrary to normalized first order derivative profiles, as in the original formulation [Cootes and Taylor, 1995, 1999, and 2001]. A nonlinear kNN-classifier is used, instead of the linear Mahalanobis distance, to find optimal displacements for landmarks. For each of the landmarks that describe the shape, at each resolution level taken into account during the segmentation optimization procedure, a distinct set of optimal features is determined. The selection of features is automatic, using the training images and sequential feature forward and backward selection. The new approach is tested on synthetic data and in four medical segmentation tasks: segmenting the right and left lung fields in a database of 230 chest radiographs, and segmenting the cerebellum and corpus callosum in a database of 90 slices from MRI brain images. In all cases, the new method produces significantly better results in terms of an overlap error measure (p < 0.001 using a paired T-test) than the original active shape model scheme.

Measuring size and shape of the hippocampus in MR images using a deformable shape model

A method for segmentation and quantification of the shape and size of the hippocampus is proposed, based on an automated image analysis algorithm. The algorithm uses a deformable shape model to locate the hippocampus in magnetic resonance images and to determine a geometric representation of its boundary. The deformable model combines three types of information. First, it employs information about the geometric properties of the hippocampal boundary, from a local and relatively finer scale to a more global and relatively coarser scale. Second, the model includes a statistical characterization of normal shape variation across individuals, serving as prior knowledge to the algorithm. Third, the algorithm utilizes a number of manually defined boundary points, which can help guide the model deformation to the appropriate boundaries, wherever these boundaries are weak or not clearly defined in MR images. Excellent agreement is demonstrated between the algorithm and manual segmentations by well-trained raters, with a correlation coefficient equal to 0.97 and algorithm/rater differences statistically equivalent to interrater differences for manual definitions.

Automatic volumetric measurement of lateral ventricles on magnetic resonance images with correction of partial volume effects

PURPOSE

To propose a method for the quantification of lateral ventricle (LV) volumes on a single sequence of 3D magnetic resonance (MR) images.

MATERIALS AND METHODS

This algorithm, following a preliminary fuzzy tissue classification step, is based on the development of mathematical morphology processes allowing both the extraction of the LVs and the correction of partial volume effects on their boundaries. The procedure is fast and totally unsupervised. The method is tested on a phantom image, then applied to five patients diagnosed as potentially suffering from Alzheimer's disease, and finally applied on several MR acquisitions to show the genericness of the algorithm.

RESULTS AND CONCLUSION

This technique yielded both an accurate estimation of ventricular volumes intra- and intersubject with respect to published data and a relevant management of partial volume effects. Numerous clinical applications are now expected, from the study of schizophrenia to the longitudinal follow-up of Alzheimer's patients.

Volume delineation by fusion of fuzzy sets obtained from multiplanar tomographic images

Techniques of three-dimensional (3-D) volume delineation from tomographic medical imaging are usually based on 2-D contour definition. For a given structure, several different contours can be obtained depending on the segmentation method used or the user's choice. The goal of this work is to develop a new method that reduces the inaccuracies generally observed. A minimum volume that is certain to be included in the volume concerned (membership degree mu = 1), and a maximum volume outside which no part of the volume is expected to be found (membership degree mu = 0), are defined semi-automatically. The intermediate fuzziness region (0 < mu < 1) is processed using the theory of possibility. The resulting fuzzy volume is obtained after data fusion from multiplanar slices. The influence of the contrast-to-noise ratio was tested on simulated images. The influence of slice thickness as well as the accuracy of the method were studied on phantoms. The absolute volume error was less than 2% for phantom volumes of 2-8 cm3, whereas the values obtained with conventional methods were much larger than the actual volumes. Clinical experiments were conducted, and the fuzzy logic method gave a volume lower than that obtained with the conventional method. Our fuzzy logic method allows volumes to be determined with better accuracy and reproducibility.

A joint physics-based statistical deformable model for multimodal brain image analysis

A probabilistic deformable model for the representation of multiple brain structures is described. The statistically learned deformable model represents the relative location of different anatomical surfaces in brain magnetic resonance images (MRIs) and accommodates their significant variability across different individuals. The surfaces of each anatomical structure are parameterized by the amplitudes of the vibration modes of a deformable spherical mesh. For a given MRI in the training set, a vector containing the largest vibration modes describing the different deformable surfaces is created. This random vector is statistically constrained by retaining the most significant variation modes of its Karhunen-Loève expansion on the training population. By these means, the conjunction of surfaces are deformed according to the anatomical variability observed in the training set. Two applications of the joint probabilistic deformable model are presented: isolation of the brain from MRI using the probabilistic constraints embedded in the model and deformable model-based registration of three-dimensional multimodal (magnetic resonance/single photon emission computed tomography) brain images without removing nonbrain structures. The multi-object deformable model may be considered as a first step toward the development of a general purpose probabilistic anatomical atlas of the brain.

Automatic segmentation of subcortical brain structures in MR images using information fusion

This paper reports a new automated method for the segmentation of internal cerebral structures using an information fusion technique. The information is provided both by images and expert knowledge, and consists in morphological, topological, and tissue constitution data. All this ambiguous, complementary and redundant information is managed using a three-step fusion scheme based on fuzzy logic. The information is first modeled into a common theoretical frame managing its imprecision and incertitude. The models are then fused and a decision is taken in order to reduce the imprecision and to increase the certainty in the location of the structures. The whole process is illustrated on the segmentation of thalamus, putamen, and head of the caudate nucleus from expert knowledge and magnetic resonance images, in a protocol involving 14 healthy volunteers. The quantitative validation is achieved by comparing computed, manually segmented structures and published data by means of indexes assessing the accuracy of volume estimation and spatial location. Results suggest a consistent volume estimation with respect to the expert quantification and published data, and a high spatial similarity of the segmented and computed structures. This method is generic and applicable to any structure that can be defined by expert knowledge and morphological images.

Multistage hybrid active appearance model matching: segmentation of left and right ventricles in cardiac MR images

A fully automated approach to segmentation of the left and right cardiac ventricles from magnetic resonance (MR) images is reported. A novel multistage hybrid appearance model methodology is presented in which a hybrid active shape model/active appearance model (AAM) stage helps avoid local minima of the matching function. This yields an overall more favorable matching result. An automated initialization method is introduced making the approach fully automated. Our method was trained in a set of 102 MR images and tested in a separate set of 60 images. In all testing cases, the matching resulted in a visually plausible and accurate mapping of the model to the image data. Average signed border positioning errors did not exceed 0.3 mm in any of the three determined contours-left-ventricular (LV) epicardium, LV and right-ventricular (RV) endocardium. The area measurements derived from the three contours correlated well with the independent standard (r = 0.96, 0.96, 0.90), with slopes and intercepts of the regression lines close to one and zero, respectively. Testing the reproducibility of the method demonstrated an unbiased performance with small range of error as assessed via Bland-Altman statistic. In direct border positioning error comparison, the multistage method significantly outperformed the conventional AAM (p < 0.001). The developed method promises to facilitate fully automated quantitative analysis of LV and RV morphology and function in clinical setting.

An adaptive-focus statistical shape model for segmentation and shape modeling of 3-D brain structures

This paper presents a deformable model for automatically segmenting brain structures from volumetric magnetic resonance (MR) images and obtaining point correspondences, using geometric and statistical information in a hierarchical scheme. Geometric information is embedded into the model via a set of affine-invariant attribute vectors, each of which characterizes the geometric structure around a point of the model from a local to a global scale. The attribute vectors, in conjunction with the deformation mechanism of the model, warranty that the model not only deforms to nearby edges, as is customary in most deformable surface models, but also that it determines point correspondences based on geometric similarity at different scales. The proposed model is adaptive in that it initially focuses on the most reliable structures of interest, and gradually shifts focus to other structures as those become closer to their respective targets and, therefore, more reliable. The proposed techniques have been used to segment boundaries of the ventricles, the caudate nucleus, and the lenticular nucleus from volumetric MR images.

Three-dimensional segmentation of anatomical structures in MR images on large data bases

In this paper an image-based method founded on mathematical morphology is presented in order to facilitate the segmentation of cerebral structures over large data bases of 3D magnetic resonance images (MRIs). The segmentation is described as an immersion simulation, applied to the modified gradient image, modeled by a generated 3D-region adjacency graph (RAG). The segmentation relies on two main processes: homotopy modification and contour decision. The first one is achieved by a marker extraction stage where homogeneous 3D-regions are identified. This stage uses contrasted regions from morphological reconstruction and labeled flat regions constrained by the RAG. Then, the decision stage intends to precisely locate the contours of regions detected by the marker extraction. This decision is performed by a 3D extension of the watershed transform. The method has been applied on a data base of 3D brain MRIs composed of fifty patients. Results are illustrated by segmenting the ventricles, corpus callosum, cerebellum, hippocampus, pons, medulla and midbrain on our data base and the approach is validated on two phantom 3D MRIs.

Object localization and border detection criteria design in edge-based image segmentation: automated learning from examples

This paper provides methodology for fully automated model-based image segmentation. All information necessary to perform image segmentation is automatically derived from a training set that is presented in a form of segmentation examples. The training set is used to construct two models representing the objects--shape model and border appearance model. A two-step approach to image segmentation is reported. In the first step, an approximate location of the object of interest is determined. In the second step, accurate border segmentation is performed. The shape-variant Hough transform method was developed that provides robust object localization automatically. It finds objects of arbitrary shape, rotation, or scaling and can handle object variability. The border appearance model was developed to automatically design cost functions that can be used in the segmentation criteria of edge-based segmentation methods. Our method was tested in five different segmentation tasks that included 489 objects to be segmented. The final segmentation was compared to manually defined borders with good results [rms errors in pixels: 1.2 (cerebellum), 1.1 (corpus callosum), 1.5 (vertebrae), 1.4 (epicardial), and 1.6 (endocardial) borders]. Two major problems of the state-of-the-art edge-based image segmentation algorithms were addressed: strong dependency on a close-to-target initialization, and necessity for manual redesign of segmentation criteria whenever new segmentation problem is encountered.

A cooperative framework for segmentation of MRI brain scans

Automatic segmentation of MRI brain scans is a complex task for two main reasons: the large variability of the human brain anatomy, which limits the use of general knowledge and, inherent to MRI acquisition, the artifacts present in the images that are difficult to process. To tackle these difficulties, we propose to mix, in a cooperative framework, several types of information and knowledge provided and used by complementary individual systems: presently, a multi-agent system, a deformable model and an edge detector. The outcome is a cooperative segmentation performed by a set of region and edge agents constrained automatically and dynamically by both, the specific gray levels in the considered image, statistical models of the brain structures and general knowledge about MRI brain scans. Interactions between the individual systems follow three modes of cooperation: integrative, augmentative and confrontational cooperation, combined during the three steps of the segmentation process namely, the specialization of the seeded-region-growing agents, the fusion of heterogeneous information and the retroaction over slices. The described cooperative framework allows the dynamic adaptation of the segmentation process to the own characteristics of each MRI brain scan. Its evaluation using realistic brain phantoms is reported.

Measurement of brain structures with artificial neural networks: two- and three-dimensional applications

PURPOSE

To evaluate the ability of an artificial neural network (ANN) to identify brain structures. This ANN was applied to postprocessed magnetic resonance (MR) images to segment various brain structures in both two- and three-dimensional applications.

MATERIALS AND METHODS

An ANN was designed that learned from experience to define the corpus callosum, whole brain, caudate, and putamen. Manual segmentation was used as a training set for the ANN. The ANN was trained on two-thirds of the manually segmented images and was tested on the remaining one-third. The reliability of the ANN was compared against manual segmentations by two technicians.

RESULTS

The ANN was able to identify the brain structures as readily and as well as did the two technicians. Reliability of the ANN compared with the technicians was 0.96 for the corpus callosum, 0.95 for the whole brain, 0.86 (right) and 0.93 (left) for the caudate, and 0.71 (right) and 0.88 (left) for the putamen.

CONCLUSION

The ANN was able to identify the structures used in this study as well as did the two technicians. The ANN could do this much more rapidly and without rater drift. Several other cortical and subcortical structures could also be readily identified with this method.