Proposed Segmentation Approach Research Articles

Automatic segmentation approach to extracting neonatal cerebral ventricles from 3D ultrasound images.

Preterm neonates with a very low birth weight of less than 1,500g are at increased risk for developing intraventricular hemorrhage (IVH). Progressive ventricle dilatation of IVH patients may cause increased intracranial pressure, leading to neurological damage, such as neurodevelopmental delay and cerebral palsy. The technique of 3D ultrasound (US) imaging has been used to quantitatively monitor the ventricular volume in IVH neonates, which may elucidate the ambiguity surrounding the timing of interventions in these patients as 2D clinical US imaging relies on linear measurement and visual estimation of ventricular dilation from a series of 2D slices. To translate 3D US imaging into the clinical setting, a fully automated segmentation algorithm is necessary to extract the ventricular system from 3D neonatal brain US images. In this paper, an automatic segmentation approach is proposed to delineate lateral ventricles of preterm neonates from 3D US images. The proposed segmentation approach makes use of phase congruency map, multi-atlas initialization technique, atlas selection strategy, and a multiphase geodesic level-sets (MGLS) evolution combined with a spatial shape prior derived from multiple pre-segmented atlases. Experimental results using 30 IVH patient images show that the proposed GPU-implemented approach is accurate in terms of the Dice similarity coefficient (DSC), the mean absolute surface distance (MAD), and maximum absolute surface distance (MAXD). To the best of our knowledge, this paper reports the first study on automatic segmentation of the ventricular system of premature neonatal brains from 3D US images.

Automatic detection of aorto-femoral vessel trajectory from whole-body computed tomography angiography data sets.

Extraction of the aorto-femoral vessel trajectory is important to utilize computed tomography angiography (CTA) in an integrated workflow of the image-guided work-up prior to trans-catheter aortic valve replacement (TAVR). The aim of this study was to develop a new, fully-automated technique for the extraction of the entire arterial access route from the femoral artery to the aortic root. An automatic vessel tracking algorithm was first used to find the centerline that connected the femoral accessing points and the aortic root. Subsequently, a deformable 3D-model fitting method was used to delineate the lumen boundary of the vascular trajectory in the whole-body CTA dataset. A validation was carried out by comparing the automatically obtained results with semi-automatically obtained results from two experienced observers. The whole framework was validated on whole body CTA datasets of 36 patients. The average Dice similarity indexes between the segmentations of the automatic method and observer 1 for the left ilio-femoral artery, the right ilio-femoral artery and the aorta were 0.977 ± 0.030, 0.980 ± 0.019, 0.982 ± 0.016; the average Dice similarity indexes between the segmentations of the automatic method and observer 2 were 0.950 ± 0.040, 0.954 ± 0.031 and 0.965 ± 0.019, respectively. The inter-observer variability resulted in a Dice similarity index of 0.954 ± 0.038, 0.952 ± 0.031 and 0.969 ± 0.018 for the left ilio-femoral artery, the right ilio-femoral artery and the aorta. The average minimal luminal diameters (MLDs) of the ilio-femoral artery were 6.03 ± 1.48, 5.70 ± 1.43 and 5.52 ± 1.32mm for the automatic method, observer 1 and observer 2 respectively. The MLDs of the aorta were 13.43 ± 2.54, 12.40 ± 2.93 and 12.08 ± 2.40mm for the automatic method, observer 1 and observer 2 respectively. The automatic measurement overestimated the MLD slightly in the ilio-femoral artery at the average by 0.323mm (SD = 0.49mm, p < 0.001) compared to observer 1 and by 0.51mm (SD = 0.71mm, p < 0.001) compared to observer 2. The proposed segmentation approach can automatically provide reliable measurements of the entire arterial accessing route that can be used to support TAVR procedures. To the best of our knowledges, this approach is the first fully automatic segmentation method of the whole aorto-femoral vessel trajectory in CTA images.

Rough-probabilistic clustering and hidden Markov random field model for segmentation of HEp-2 cell and brain MR images

The segmentation of images into different meaningful classes is an important task for automatic image analysis technique. The finite Gaussian mixture model is one of the popular models for parametric model based image segmentation. However, the normality assumption of this model induces certain limitations as a single representative value is considered to represent each class. In this regard, the paper presents a new clustering algorithm, termed as rough-probabilistic clustering, integrating judiciously the merits of rough sets and a new probability distribution, called stomped normal (SN) distribution. The intensity distribution of a class is represented by SN distribution, where each class consists of a crisp lower approximation and a probabilistic boundary region. The intensity distribution of any image is modeled as a mixture of finite number of SN distributions. The expectation–maximization algorithm is used to estimate the parameters of each class. Incorporating hidden Markov random field framework into rough-probabilistic clustering, a new method is proposed for accurate and robust segmentation of images. The performance of the proposed segmentation approach, along with a comparison with related methods, is demonstrated on a set of HEp-2 cell images, and synthetic and real brain MR images for different bias fields and noise levels.

Color-based automatic quality control for roasting chicken

In this paper, we develop a color-based automatic control system for roasting chicken industry by using the modern Combi Oven and new image processing techniques. Firstly, we embedded a video camera into the Combi Oven and design a color monitoring system. Secondly, a new segmentation technique based on level set approach is developed. Thirdly, a new color indicator is defined for color quality control. Finally, extensive experiments have been conducted to validate the proposed segmentation approach and the effectiveness of the developed system.

Delineation and diagnosis of brain tumors from post contrast T1-weighted MR images using rough granular computing and random forest

This paper presents a new approach of delineation and characterization of four different types of brain tumors viz. Glioblastoma multiforme (GBM), metastasis (MET), meningioma (MG) and granuloma (GN) from magnetic resonance imaging (MRI) slices of post contrast T1-weighted (T1C) sequence to improve the computer assistive diagnostic accuracy. An integrated framework of identification and extraction of tumor region, quantification of histogram, shape and textural features followed through pattern classification by machine learning algorithm has been proposed. Rough entropy based thresholding in granular computing paradigm has been adopted for delineation of tumor area. After accomplishing quantitative validation and comparison with existing methods, experimental results prove the efficiency and applicability of proposed segmentation approach. In the next stage, the extracted lesions have been quantified with 86 features to develop the training dataset. Random forest (RF), an ensemble learning scheme has been implemented, which learns the training data for accurate prediction of the class label of a given input. The performance of RF has been evaluated by statistical measures from 3 fold cross-validation and compared with five different classifiers. The same experiment has been repeated over the reduced set of features generated by correlation based feature selection strategy. Experimental results show the superiority of RF (Sensitivity achieved in %: GBM-96.7, MET-96.2, MG-98.1 and GN-97.7) with the complete set of features. The comparison of proposed methodology with the existing works signifies its applicability and effectiveness. Additionally a 10 fold cross-validation has been accomplished to justify the statistical significance of the classification accuracy achieved from proposed methodology.

Myocardial Iron Loading Assessment by Automatic Left Ventricle Segmentation with Morphological Operations and Geodesic Active Contour on T2* images.

Myocardial iron loading thalassemia patients could be identified using T2* magnetic resonance images (MRI). To quantitatively assess cardiac iron loading, we proposed an effective algorithm to segment aligned free induction decay sequential myocardium images based on morphological operations and geodesic active contour (GAC). Nine patients with thalassemia major were recruited (10 male and 16 female) to undergo a thoracic MRI scan in the short axis view. Free induction decay images were registered for T2* mapping. The GAC were utilized to segment aligned MR images with a robust initialization. Segmented myocardium regions were divided into sectors for a region-based quantification of cardiac iron loading. Our proposed automatic segmentation approach achieve a true positive rate at 84.6% and false positive rate at 53.8%. The area difference between manual and automatic segmentation was 25.5% after 1000 iterations. Results from T2* analysis indicated that regions with intensity lower than 20 ms were suffered from heavy iron loading in thalassemia major patients. The proposed method benefited from abundant edge information of the free induction decay sequential MRI. Experiment results demonstrated that the proposed method is feasible in myocardium segmentation and was clinically applicable to measure myocardium iron loading.

Integrating multi-criteria decision making and clustering for business customer segmentation

Purpose – The purpose of this paper is to develop a systematic approach for business customer segmentation. Design/methodology/approach – This study proposes an approach for business customer segmentation that integrates clustering and multi-criteria decision making (MCDM). First, proper segmentation variables are identified and then customers are grouped by using hierarchical and partitional clustering algorithms. The approach extended the recency-frequency-monetary (RFM) model by proposing five novel segmentation variables for business markets. To confirm the viability of the proposed approach, a real-world application is presented. Three agglomerative hierarchical clustering algorithms namely “Ward’s method,” “single linkage” and “complete linkage,” and a partitional clustering algorithm, “k-means,” are used in segmentation. In the implementation, fuzzy analytic hierarchy process is employed to determine the importance of the segments. Findings – Business customers of an international original equipment manufacturer (OEM) are segmented in the application. In this regard, 317 business customers of the OEM are segmented as “best,” “valuable,” “average,” “potential valuable” and “potential invaluable” according to the cluster ranks obtained in this study. The results of the application reveal that the proposed approach can effectively be used in practice for business customer segmentation. Research limitations/implications – The success of the proposed approach relies on the availability and quality of customers’ data. Therefore, design of an extensive customer database management system is the foundation for any successful customer relationship management (CRM) solution offered by the proposed approach. Such a database management system may entail a noteworthy level of investment. Practical implications – The results of the application reveal that the proposed approach can effectively be used in practice for business customer segmentation. By making customer segmentation decisions, the proposed approach can provides firms a basis for the development of effective loyalty programs and design of customized strategies for their customers. Social implications – The proposed segmentation approach may contribute firms to gaining sustainable competitive advantage in the market by increasing the effectiveness of CRM strategies. Originality/value – This study proposes an integrated approach for business customer segmentation. The proposed approach differentiates itself from its counterparts by combining MCDM and clustering in business customer segmentation. In addition, it extends the traditional RFM model by including five novel segmentation variables for business markets.

An Accurate Approach for Automatic Segmentation of Power Distribution Voltage Waveforms

This paper addresses one of the fundamental steps in automatic waveform analysis: transient segmentation. We present a new approach which incorporates the advantages of a multilevel wavelet decomposition and the representation of the support vector data description. Real data from a monitoring system developed for lightning overvoltage detection in overhead distribution power lines was used for comparison and validation of segmentation performance. The experiments involve the proposed segmentation approach and usual segmentation methods, such as Kalman filtering, autoregressive models, and standard discrete wavelet transform. The results show that the proposed segmentation method based on DWT+SVDD yields better overall accuracy for transient segmentation when compared to currently used methods, demonstrating the potential for applications in oscillographic recorders for smart distribution networks, where identification, characterization, and mitigation of events are critical for network operation and maintenance.

Unsupervised segmentation of highly dynamic scenes through global optimization of multiscale cues

We propose a novel method for highly dynamic scene segmentation by formulating foreground object extraction as a global optimization framework that integrates a set of multiscale spatio-temporal cues. The multiscale features consist of a combination of motion and spectral components at a pixel level as well as spatio-temporal consistency constraints between superpixels. To compensate for the ambiguities of foreground hypothesis due to highly dynamic and cluttered backgrounds, we formulate salient foreground mapping as a convex optimization of weighted total variation energy, which is efficiently solved by using an alternating minimization scheme. Moreover, the appearance and position spatio-temporal consistency constraints between superpixels are explicitly incorporated into a Markov random field energy functional for further refinement of the set of salient pixel-level foreground mapping. This work facilitates sequential integration of multiscale probability constraints into a global optimal segmentation framework to help address object boundary ambiguities in the case of highly dynamic scenes. Extensive experiments on challenging dynamic scene data sets demonstrate the feasibility and superiority of the proposed segmentation approach.

Automated Segmentation of Optical Nerves by Neural Network based Region Growing

Computer Aided Diagnosis (CAD) of retinal image has been a revolutionary step in the early diagnosis of diseases present in the eye. Developing an efficient and robust algorithm for optical nerve segmentation has been a demanding area of growing research of interest during the last two decades. The initial step in computer aided diagnosis of retinal image is generally to segment the nerves present in it and then to analyze each area separately in order to find the presence of pathologies present in it. This research reports on segmentation of the nerves by segmenting the retinal images using Echo State Neural Networks along with the combination of region growing algorithm. Region growing has been combined with ESNN in this work since it reduces the number of steps in segmentation for the process of identifying a tissue in the CT retinal image. The performance of this proposed segmentation is proved to be better when it is compared with other existing conventional segmentation algorithms. From the experimental results, it has been observed that the proposed segmentation approach provides better segmentation accuracy.

User-Guided Segmentation of Preterm Neonate Ventricular System from 3-D Ultrasound Images Using Convex Optimization

A three-dimensional (3-D) ultrasound (US) system has been developed to monitor the intracranial ventricular system of preterm neonates with intraventricular hemorrhage (IVH) and the resultant dilation of the ventricles (ventriculomegaly). To measure ventricular volume from 3-D US images, a semi-automatic convex optimization-based approach is proposed for segmentation of the cerebral ventricular system in preterm neonates with IVH from 3-D US images. The proposed semi-automatic segmentation method makes use of the convex optimization technique supervised by user-initialized information. Experiments using 58 patient 3-D US images reveal that our proposed approach yielded a mean Dice similarity coefficient of 78.2% compared with the surfaces that were manually contoured, suggesting good agreement between these two segmentations. Additional metrics, the mean absolute distance of 0.65 mm and the maximum absolute distance of 3.2 mm, indicated small distance errors for a voxel spacing of 0.22 × 0.22 × 0.22 mm3. The Pearson correlation coefficient (r = 0.97, p < 0.001) indicated a significant correlation of algorithm-generated ventricular system volume (VSV) with the manually generated VSV. The calculated minimal detectable difference in ventricular volume change indicated that the proposed segmentation approach with 3-D US images is capable of detecting a VSV difference of 6.5 cm3 with 95% confidence, suggesting that this approach might be used for monitoring IVH patients' ventricular changes using 3-D US imaging. The mean segmentation times of the graphics processing unit (GPU)- and central processing unit-implemented algorithms were 50 ± 2 and 205 ± 5 s for one 3-D US image, respectively, in addition to 120 ± 10 s for initialization, less than the approximately 35 min required by manual segmentation. In addition, repeatability experiments indicated that the intra-observer variability ranges from 6.5% to 7.5%, and the inter-observer variability is 8.5% in terms of the coefficient of variation of the Dice similarity coefficient. The intra-class correlation coefficient for ventricular system volume measurements for each independent observer ranged from 0.988 to 0.996 and was 0.945 for three different observers. The coefficient of variation and intra-class correlation coefficient revealed that the intra- and inter-observer variability of the proposed approach introduced by the user initialization was small, indicating good reproducibility, independent of different users.

A multi-atlas-based segmentation framework for prostate brachytherapy.

Low-dose-rate brachytherapy is a radiation treatment method for localized prostate cancer. The standard of care for this treatment procedure is to acquire transrectal ultrasound images of the prostate in order to devise a plan to deliver sufficient radiation dose to the cancerous tissue. Brachytherapy planning involves delineation of contours in these images, which closely follow the prostate boundary, i.e., clinical target volume. This process is currently performed either manually or semi-automatically, which requires user interaction for landmark initialization. In this paper, we propose a multi-atlas fusion framework to automatically delineate the clinical target volume in ultrasound images. A dataset of a priori segmented ultrasound images, i.e., atlases, is registered to a target image. We introduce a pairwise atlas agreement factor that combines an image-similarity metric and similarity between a priori segmented contours. This factor is used in an atlas selection algorithm to prune the dataset before combining the atlas contours to produce a consensus segmentation. We evaluate the proposed segmentation approach on a set of 280 transrectal prostate volume studies. The proposed method produces segmentation results that are within the range of observer variability when compared to a semi-automatic segmentation technique that is routinely used in our cancer clinic.

Recognition of gait cycle phases using wearable sensors

The analysis and monitoring of the human daily living activities plays an important role for rehabilitation goals, fall prevention rehabilitation and general health-care treatments. Among these activities, walking is the most important daily motion. Studying the evolution of the gait cycle through the analysis of the human center of force is beneficial to predict any abnormal walking pattern. The analysis is based on the use of pressure-based mapping system that collects pressure and force measurement during the gait cycle. This paper deals mainly with the detection of the main characteristics of the gait phases. To this end, a segmentation of the center of force of the human body measure through the in-shoe pressure mapping system is performed to identify the gait phases. The proposed segmentation approach consists in modeling each segment by a regression model and using logistic functions to model the transitions between segments. This flexible modeling through the logistic functions has the advantage of detecting abrupt and smooth transitions between segments.

A novel technique to extract accurate cell contours applied for segmentation of phytoplankton images

Active contour model (ACM) is an image segmentation technique widely applied for object detection. Most of the research in ACM area is dedicated to the development of various energy functions based on physical intuition. Here, instead of constructing a new energy function, we manipulate values of ACM parameters to generate a multitude of potential contours, score them using a machine-learned ranking technique, and select the best contour for each object in question. Several learning-to-rank (L2R) methods are evaluated with a goal to choose the most accurate in assessing the quality of generated contours. Superiority of the proposed segmentation approach over the original boosted edge-based ACM and three ACM implementations using the level-set framework is demonstrated for the task of Prorocentrum minimum cells' detection in phytoplankton images. Experiments show that diverse set of contour features with grading learned by a variant of multiple additive regression trees ($$\uplambda $$?-MART) helped to extract precise contour for 87.6 % of cells tested.

Trainable COSFIRE filters for vessel delineation with application to retinal images.

Retinal imaging provides a non-invasive opportunity for the diagnosis of several medical pathologies. The automatic segmentation of the vessel tree is an important pre-processing step which facilitates subsequent automatic processes that contribute to such diagnosis. We introduce a novel method for the automatic segmentation of vessel trees in retinal fundus images. We propose a filter that selectively responds to vessels and that we call B-COSFIRE with B standing for bar which is an abstraction for a vessel. It is based on the existing COSFIRE (Combination Of Shifted Filter Responses) approach. A B-COSFIRE filter achieves orientation selectivity by computing the weighted geometric mean of the output of a pool of Difference-of-Gaussians filters, whose supports are aligned in a collinear manner. It achieves rotation invariance efficiently by simple shifting operations. The proposed filter is versatile as its selectivity is determined from any given vessel-like prototype pattern in an automatic configuration process. We configure two B-COSFIRE filters, namely symmetric and asymmetric, that are selective for bars and bar-endings, respectively. We achieve vessel segmentation by summing up the responses of the two rotation-invariant B-COSFIRE filters followed by thresholding. The results that we achieve on three publicly available data sets (DRIVE: Se=0.7655, Sp=0.9704; STARE: Se=0.7716, Sp=0.9701; CHASE_DB1: Se=0.7585, Sp=0.9587) are higher than many of the state-of-the-art methods. The proposed segmentation approach is also very efficient with a time complexity that is significantly lower than existing methods.

A robust MR-based attenuation map generation in short-TE MR images of the head employing hybrid spatial fuzzy C-means clustering and intensity inhomogeneity correction.

Deriving an accurate attenuation correction map (μ-map) from magnetic resonance (MR) volumes has become an important problem in hybrid PET/MR imaging. Recently, short echo-time (STE) MR imaging technique incorporating fuzzy C-means (FCM) tissue classification and 2-point Dixon image acquisition has been introduced as a feasible approach for segmentation of the bone from air and soft tissue. However, this method imposes additional imaging and the performance of the standard FCM algorithm, suffering from the lack of spatial information, becomes impaired in the presence of inherent noise and intensity inhomogeneity. Here, we exploit a spatial fuzzy C-means (SFCM) segmentation algorithm in combination with a robust intensity inhomogeneity correction method on single STE-MR images, to differentiate various tissue classes. MR images of five subjects were acquired on a clinical 1.5T MRI System, MAGNETOM Avanto, using a FLASH 3D pulse sequence with TE=1.1ms, TR=12ms, flip angle=18°, voxel size=1.2×1.2×2mm3. The proposed segmentation approach consists of four main steps: (1) Intensity-inhomogeneity correction, to separate air and bone in the regions with high inhomogeneity, like the nasal areas; (2) applying SFCM to segment the images into four clusters including air, a part of soft tissue and bone, and two other soft tissue classes. Upon this step, the air cluster would be accurately separated; (3) employing shape factor analysis to remove the eyes with close signal intensity to that of bone; and (4) μ-map generation by downsampling and assigning attenuation coefficients to the corresponding segmented tissues. Quantitative evaluation indicated sensitivity of over 95% for air and soft tissue and about 81% in the bony region, and specificity of over 95% for all tissue classes. The proposed STE-MR imaging in combination with the segmentation technique could be potentially exploited as an efficient approach to generate MR-based attenuation correction maps in clinical PET/MR applications, where bone and air coexist.

Partial volume correction of brain perfusion estimates using the inherent signal data of time-resolved arterial spin labeling.

Quantitative perfusion MRI based on arterial spin labeling (ASL) is hampered by partial volume effects (PVEs), arising due to voxel signal cross-contamination between different compartments. To address this issue, several partial volume correction (PVC) methods have been presented. Most previous methods rely on segmentation of a high-resolution T1 -weighted morphological image volume that is coregistered to the low-resolution ASL data, making the result sensitive to errors in the segmentation and coregistration. In this work, we present a methodology for partial volume estimation and correction, using only low-resolution ASL data acquired with the QUASAR sequence. The methodology consists of a T1 -based segmentation method, with no spatial priors, and a modified PVC method based on linear regression. The presented approach thus avoids prior assumptions about the spatial distribution of brain compartments, while also avoiding coregistration between different image volumes. Simulations based on a digital phantom as well as in vivo measurements in 10 volunteers were used to assess the performance of the proposed segmentation approach. The simulation results indicated that QUASAR data can be used for robust partial volume estimation, and this was confirmed by the in vivo experiments. The proposed PVC method yielded probable perfusion maps, comparable to a reference method based on segmentation of a high-resolution morphological scan. Corrected gray matter (GM) perfusion was 47% higher than uncorrected values, suggesting a significant amount of PVEs in the data. Whereas the reference method failed to completely eliminate the dependence of perfusion estimates on the volume fraction, the novel approach produced GM perfusion values independent of GM volume fraction. The intra-subject coefficient of variation of corrected perfusion values was lowest for the proposed PVC method. As shown in this work, low-resolution partial volume estimation in connection with ASL perfusion estimation is feasible, and provides a promising tool for decoupling perfusion and tissue volume.

Online Persian/Arabic script classification without contextual information

Segmentation accuracy plays a vital role in script recognition process and therefore, research in this area is still fresh. Accordingly, this paper presents a new method for segmenting cursive handwritten Persian/Arabic words. Moreover, this paper proposes a novel strategy to tackle the challenge of recognising the handwritten sub-word as input without contextual information. The recognition strategy is tested with Persian scripts but can be used for both Person and Arabic scripts. Initially, the input data are pre-processed to generate desired data. Later, the proposed segmentation approach fragments the pre-processed data at regular intervals. The discriminative features are extracted as a binary input set for a hybrid of artificial neural networks and particle swarm optimisation (ANN–PSO) classifier to recognize Persian script. Finally, the achieved segmentation and recognition rates are compared with the results reported in the literature. Segmentation and recognition rates thus achieved exhibit promising achievements in the state of art.

Intelligent Approach for Segmenting CT Lung Images Using Fuzzy Logic with Bitplane

In this article, we present a new grey scale image segmentation method based on Fuzzy logic and bitplane techniques which combines the bits of different bitplanes of a pixel inorder to increase the segmentation quality and to get a more reliable and accurate segmentation result. The proposed segmentation approach is conceptually different and explores a new strategy. Infact, our technique consists in combining many realizations of the image together inorder to increase the information quality and to get an optimal segmented image. For segmentation, we proceed in two steps. In the first step, we begin by identifying the bitplanes that represent the lungs clearly. For this purpose, the intensity value of a pixel is separated into bitplanes. In the second step, segmentation values are assigned for each bitplane based on membership table. The segmented values of foreground are combined and the segmentation values of background are combined. The algorithm is demonstrated through the medical computed tomography (CT) images. The segmentation accuracy of the proposed method is compared with two existing techniques. Satisfactory segmentation results have been obtained showing the effectiveness and superiority of the proposed method.

Vertebral body segmentation with prior shape constraints for accurate BMD measurements

We propose a novel vertebral body segmentation approach, which is based on the graph cuts technique with shape constraints. The proposed approach depends on both image appearance and shape information. Shape information is gathered from a set of training shapes. Then we estimate the shape variations using a new distance probabilistic model which approximates the marginal densities of the vertebral body and its background in the variability region using a Poisson distribution refined by positive and negative Gaussian components. To segment a vertebral body, we align its 3D shape with the training 3D shape so we can use the distance probabilistic model. Then its gray level is approximated with a Linear Combination of Gaussians (LCG) with sign-alternate components. The spatial interaction between the neighboring voxels is identified using a new analytical approach. Finally, we formulate an energy function using both appearance models and shape constraints. This function is globally minimized using s/t graph cuts to get the optimal segmentation. Experimental results show that the proposed technique gives promising results compared to other alternatives. Applications on Bone Mineral Density (BMD) measurements of vertebral body are given to illustrate the accuracy of the proposed segmentation approach.