Heart chambers and whole heart segmentation techniques: review

Abstract

Abstract. Computer-aided segmentation of cardiac images obtained by various modalities plays an important role and is a prerequisite for a wide range of cardiac applications by facilitating the delineation of anatomical regions of interest. Numerous computerized methods have been developed to tackle this problem. Recent studies employ sophisticated techniques using available cues from cardiac anatomy such as geometry, visual appearance, and prior knowledge. In addition, new minimization and computational methods have been adopted with improved computational speed and robustness. We provide an overview of cardiac segmentation techniques, with a goal of providing useful advice and references. In addition, we describe important clinical applications, imaging modalities, and validation methods used for cardiac segmentation.

Automatic image-based segmentation of the heart from CT scans

“…Compared with other imaging modalities (such as ultrasound and magnetic resonance imaging), cardiac computed tomography (CT) can provide detailed anatomical information about the heart chambers, great vessels, and coronary arteries [4,5]. Actually, CT is often preferred by diagnosticians since it provides more accurate anatomical information about the visualized structures, thanks to its higher signal-to-noise ratio and better spatial resolution.…”

“…These works deal with different strategies for approaching the segmentation task, including image-driven algorithms [10][11][12][13], probabilistic atlases [14,15], fuzzy clustering [16], deformable models [17][18][19], neural networks [20], active appearance models [21,22], anatomical-based landmarks [23], or level set and its variations [24,25]. A comprehensive review of techniques commonly used in cardiac image segmentation can be found in Kang et al [5]. Nevertheless, many published methods have various disadvantages for routine clinical practice: they are either computationally demanding [6,14,16,22], potentially unstable for subjects with pathology [25,26], limited to the left ventricle [11,24,25,27], require additional images to be acquired [28,29], or need complex shape and/or gray-level appearance models constructed (or 'learned') from many manually segmented images -which is labor intensive and of limited use due to both anatomical and image contrast inconsistencies [14,22,[26][27][28].…”

Regional ejection fraction and regional area strain for left ventricular function assessment in male patients after first-time myocardial infarction

“…Automatic image segmentation techniques [32,33] could be explored to speed up the manual segmentation process. Other segmentation techniques such as shape-based interpolation [34] and superresolution [35,36] can also be potentially used to automate and increase the accuracy of the cardiac segmentation.…”

Right Ventricle Segmentation by Temporal Information Constrained Gradient Vector Flow

“…and 1 In the literature, several approaches (e.g., [16] [17]) have been proposed to segment the cardiac images by using time constrained information. In [16], the authors introduced prior motion knowledge for the contour deformation through temporal smoothing (average of two adjacent frames) or trajectory constraints (to constrain the shape to its reference trajectory).…”

“…We refer the reader to the two recently published survey papers [1] and [2] for detailed reviews on the topic. In contrast, due to its higher anatomical complexity, the right ventricle (RV) segmentation has not been extensively explored.…”