Automated Segmentation of Left Ventricle Using Local and Global Intensity Based Active Contour and Dynamic Programming (original) (raw)
Related papers
Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, 2011
A statistical method is proposed to compare several estimates of a relevant clinical parameter when no gold standard is available. The method is illustrated by considering the left ventricle ejection fraction derived from cardiac magnetic resonance images and computed using seven approaches with different degrees of automation. The proposed method did not use any a priori regarding with the reliability of each method and its degree of automation. The results showed that the most accurate estimates of the ejection fraction were obtained using manual segmentations, followed by the semiautomatic methods, while the methods with the least user input yielded the least accurate ejection fraction estimates. These results were consistent with the expected performance of the estimation methods, suggesting that the proposed statistical approach might be helpful to assess the performance of estimation methods on clinical data for which no gold standard is available.
IEEE Transactions on Biomedical Engineering, 2000
An automatic left ventricle (LV) segmentation algorithm is presented for quantification of cardiac output and myocardial mass in clinical practice. The LV endocardium is first segmented using region growth with iterative thresholding by detecting the effusion into the surrounding myocardium and tissues. Then the epicardium is extracted using the active contour model guided by the endocardial border and the myocardial signal information estimated by iterative thresholding. This iterative thresholding and active contour model with adaptation (ITHACA) algorithm was compared to manual tracing used in clinical practice and the commercial MASS Analysis software (General Electric) in 38 patients, with Institutional Review Board (IRB) approval. The ITHACA algorithm provided substantial improvement over the MASS software in defining myocardial borders. The ITHACA algorithm agreed well with manual tracing with a mean difference of blood volume and myocardial mass being 2.9 ± 6.2 mL (mean ± standard deviation) and −0.9 ± 16.5 g, respectively. The difference was smaller than the difference between manual tracing and the MASS software (approximately −20.0 ± 6.9 mL and −1.0 ± 20.2 g, respectively). These experimental results support that the proposed ITHACA segmentation is accurate and useful for clinical practice. Index Terms-Active contour model, cardiac MRI, cine MRI, iterative thresholding, left ventricle (LV) segmentation. I. INTRODUCTION C ARDIAC disease is the leading cause of death in the USA and the developed world. The quantification of myocardial mass and systolic function is routinely performed in the clinical setting to diagnose a variety of cardiac pathologies. MRI, computed tomography (CT), ultrasound, X-ray, and single photon emission computed tomography (SPECT) can all be Manuscript
New semi-automated segmentation approach of the left ventricle applied to cine MR images analysis
International Journal of Applied Pattern Recognition, 2018
Cardiovascular abnormalities have become one of the most dangerous diseases. They affect people around the entire world and some countries are more concerned by this illness. In this paper, a new methodological approach dedicated to analyse and delineate the short-axis cardiac MRI of the left ventricle (LV) is presented. In this case, thresholding and morphological operation, active contours model and region growing are combined to extract endocardial exactly. The systole volume (VTS), diastole volume (VTD) and ejection volume (EV) are then successively calculated to predict the cardiovascular diseases. The results are validated by a database (Cousty et al., 2010) where the expert manual contouring was available and the similarity index (Jaccard index) between the proposed method and expert segmentations are calculated. After discussion, we conclude that the presented method leads to satisfying results end-diastolic endocardium (0.92) and endsystolic endocardium (0.88), achieving both fast calculation and accuracy objectives.
European Journal of Nuclear Medicine and Molecular Imaging, 1998
& p . 1 : Tomographic multi-gated blood pool scintigraphy (TMUGA) is a widely available method which permits simultaneous assessment of right and left ventricular ejection fractions. However, the widespread clinical use of this technique is impeded by the lack of segmentation methods dedicated to an automatic analysis of ventricular activities. In this study we evaluated how a watershed algorithm succeeds in providing semi-automatic segmentation of ventricular activities in order to measure right and left ejection fractions by TMUGA. The left ejection fractions of 30 patients were evaluated both with TMUGA and with planar multi-gated blood pool scintigraphy (PMUGA). Likewise, the right ejection fractions of 25 patients were evaluated with first-pass scintigraphy (FP) and with TMUGA. The watershed algorithm was applied to the reconstructed slices in order to group together the voxels whose activity came from one specific cardiac cavity. First, the results of the watershed algorithm were compared with manual drawing around left and right ventricles. Left ejection fractions evaluated by TMUGA with the watershed procedure were not significantly different (p=0.30) from manual outlines whereas a small but significant difference was found for right ejection fractions (p=0.004). Then right and left ejection fractions evaluated by TMUGA (with the semi-automatic segmentation procedure) were compared with the results obtained by FP or PMUGA. Left ventricular ejection fractions evaluated by TMUGA showed an excellent correlation with those evaluated by PMUGA (r=0.93; SEE=5.93%; slope=0.99; intercept = 4.17%). The measurements of these ejection fractions were significantly higher with TMUGA than with PMUGA (P<0.01). The interoperator variability for the measurement of left ejection fractions by TMUGA was 4.6%. Right ventricular ejection fractions evaluated by TMUGA showed a good correlation with those evaluated by FP (r = 0.81; SEE = 6.68%; slope = 1.00; intercept = 0.85%) and were not significantly different (P = 0.42).
Magnetic Resonance Materials in Physics, Biology and Medicine, 2007
Objective In the field of cardiac MR image segmentation, active contour models, or snakes, have been extensively used, owing to their promising results and to the numerous extensions proposed to improve their performance. This paper explores a methodology for evaluating cardiac MR image segmentation algorithms, which assesses the distance between computergenerated and the observer's hand-outlined boundaries. This metric was applied to various external force extensions of the traditional snake, since no systematic comparison has been performed. Materials and methods Cardiac MRI from six patients were analyzed. Imaging was performed on a 1.5 T MR scanner with ECG-gated balanced steady-state free precession (b-SSFP) sequences. Segmentation performances were established for traditional snake, gradient vector flow snake, standard-and guided-pressure forcebased snake. The use of a pre-treatment with non-linear anisotropic filtering was also compared to non-filtered images. Results Agreement between manual and segmentation algorithms was satisfactory for ejection fraction for every segmentation scheme. However end-systolic and end-diastolic volumes were systematically underestimated. Conclusion The developed regional error metric provided a more rigorous evaluation of the segmentation schemes in comparison to the classical derived parameters based on left ventricle volume estimation, usually used in functional cardiac MR studies. These derived
A Semi-Automated Method for Measurement of Left Ventricular Volumes in 3D Echocardiography
IEEE Access
Segmentation of the left ventricle in echocardiography data currently poses a challenge, where delineation of the endocardial borders is a time consuming and difficult task. Though semi-automated and fully automated methods have been developed for left ventricular segmentation, they suffer from a number of drawbacks. These drawbacks include the dependence on large sets of training data and assumptions about the distribution of the intensities of the image. This paper proposes a novel volumetric segmentation algorithm based on an angular slicing approach for 3-D echocardiography scans and a diffeomorphic nonrigid registration method. The proposed method is fast, reproducible, and yields a volumetric segmentation with minimal user interaction. The algorithm was evaluated on 30 participants from the challenge on endocardial 3-D ultrasound segmentation dataset from the medical image computing and computer assisted interventions Challenge 2014. The proposed method yielded the following average distance metrics for the end diastolic volumes: 1) mean absolute distance of 2.36 mm, 2) Hausdorff distance of 8.25 mm, and 3) Dice score of 0.887. For the end systolic volumes, the following average distance metrics were obtained: 1) mean absolute distance of 2.33 mm, 2) Hausdorff distance of 8.95 mm, and 3) Dice score of 0.857. The following clinical metrics for the ejection fraction are reported: 1) modified correlation coefficient of 0.169, 2) bias in mL of −3.96 mL, and 3) standard deviation of 6.85 mL. The results demonstrate the robustness of the proposed volumetric segmentation approach.