Fast, accurate, and fully automatic segmentation of the right ventricle in short-axis cardiac MRI

• Published in: Computerized medical imaging and graphics Vol. 38; no. 3; pp. 190 - 201
• Main Authors: Ringenberg, Jordan, Deo, Makarand, Devabhaktuni, Vijay, Berenfeld, Omer, Boyers, Pamela, Gold, Jeffrey
• Format: Journal Article
• Language: English
• Published: New York, NY Elsevier Ltd 01-04-2014; Elsevier
• Subjects: A priori constraints; Algorithms; Applied sciences; Artificial Intelligence; Binary difference of Gaussians filter; Biological and medical sciences; Cardiac MRI; Cardiovascular system; Computer science; control theory; systems; Exact sciences and technology; Heart Ventricles - pathology; Humans; Image Enhancement - methods; Image Interpretation, Computer-Assisted - methods; Internal Medicine; Investigative techniques, diagnostic techniques (general aspects); Magnetic Resonance Imaging, Cine - methods; Medical sciences; Optimal thresholding; Other; Pattern Recognition, Automated - methods; Pattern recognition. Digital image processing. Computational geometry; Radiodiagnosis. Nmr imagery. Nmr spectrometry; Reproducibility of Results; Sensitivity and Specificity; Ventricular Dysfunction, Right - pathology; Ventricular segmentation; A priori constraints; Binary difference of Gaussians filter; Ventricular segmentation; Cardiac MRI; Optimal thresholding; Measurement; Geometrical shape; Computer vision; Confinement; Gaussian filtering; Threshold detection; Right ventricle; Nuclear magnetic resonance imaging; Image segmentation; Morphology; Band pass filter; Threshold
• ISSN: 0895-6111; 1879-0771
• DOI: 10.1016/j.compmedimag.2013.12.011

Abstract This paper presents a fully automatic method to segment the right ventricle (RV) from short-axis cardiac MRI. A combination of a novel window-constrained accumulator thresholding technique, binary difference of Gaussian (DoG) filters, optimal thresholding, and morphology are utilized to drive the segmentation. A priori segmentation window constraints are incorporated to guide and refine the process, as well as to ensure appropriate area confinement of the segmentation. Training and testing were performed using a combined 48 patient datasets supplied by the organizers of the MICCAI 2012 right ventricle segmentation challenge, allowing for unbiased evaluations and benchmark comparisons. Marked improvements in speed and accuracy over the top existing methods are demonstrated.