Ahmed S. Fahmy

Fast Localization of the Optic Disc Using Projection of Image Features

Optic Disc (OD) localization is an important pre-processing step that significantly simplifies subsequent segmentation of the OD and other retinal structures. Current OD localization techniques suffer from impractically-high computation times (few minutes per image). In this work, we present a fast technique that requires less than a second to localize the OD. The technique is based upon obtaining two projections of certain image features that encode the x- and y- coordinates of the OD. The resulting 1-D projections are then searched to determine the location of the OD. This avoids searching the 2-D image space and, thus, enhances the speed of the OD localization process. Image features such as retinal vessels orientation and the OD brightness are used in the current method. Four publicly-available databases, including STARE and DRIVE, are used to evaluate the proposed technique. The OD was successfully located in 330 images out of 340 images (97%) with an average computation time of 0.65 s.

A collaborative resource to build consensus for automated left ventricular segmentation of cardiac MR images.

A collaborative framework was initiated to establish a community resource of ground truth segmentations from cardiac MRI. Multi-site, multi-vendor cardiac MRI datasets comprising 95 patients (73 men, 22 women; mean age 62.73±11.24years) with coronary artery disease and prior myocardial infarction, were randomly selected from data made available by the Cardiac Atlas Project (Fonseca et al., 2011). Three semi- and two fully-automated raters segmented the left ventricular myocardium from short-axis cardiac MR images as part of a challenge introduced at the STACOM 2011 MICCAI workshop (Suinesiaputra et al., 2012). Consensus myocardium images were generated based on the Expectation-Maximization principle implemented by the STAPLE algorithm (Warfield et al., 2004). The mean sensitivity, specificity, positive predictive and negative predictive values ranged between 0.63 and 0.85, 0.60 and 0.98, 0.56 and 0.94, and 0.83 and 0.92, respectively, against the STAPLE consensus. Spatial and temporal agreement varied in different amounts for each rater. STAPLE produced high quality consensus images if the region of interest was limited to the area of discrepancy between raters. To maintain the quality of the consensus, an objective measure based on the candidate automated rater performance distribution is proposed. The consensus segmentation based on a combination of manual and automated raters were more consistent than any particular rater, even those with manual input. The consensus is expected to improve with the addition of new automated contributions. This resource is open for future contributions, and is available as a test bed for the evaluation of new segmentation algorithms, through the Cardiac Atlas Project (www.cardiacatlas.org).

Real-time MR imaging of myocardial regional function using strain-encoding (SENC) with tissue through-plane motion tracking.

To implement real‐time myocardial strain‐encoding (SENC) imaging in combination with tracking the tissue displacement in the through‐plane direction.

Floating Navigator Echo (FNAV) for In-Plane 2D Translational Motion Estimation

A modification of the classical navigator echo (NAV) technique is presented whereby both 2D translational motion components are computed from a single navigator line. Instead of acquiring the NAV at the center of the k‐space, a kx line is acquired off‐center in the phase‐encoding (ky) direction as a floating NAV (FNAV). It is shown that the translational motion in both the readout and phase‐encoding directions can be computed from this line. The algorithm used is described in detail and verified experimentally. The new technique can be readily implemented to replace classic NAV in MRI sequences, with little to no additional cost or complexity. The new method can help suppress 2D translational motion and provide more accurate motion estimates for other motion‐suppression techniques, such as the diminishing variance algorithm. Magn Reson Med 51:403–407, 2004.

Regional right ventricular function and timing of contraction in healthy volunteers evaluated by strain‐encoded MRI

To prospectively determine the feasibility and accuracy of strain‐encoded (SENC) magnetic resonance imaging (MRI) for the characterization of the right ventricular free wall (RVFW) strain and timing of contraction at 3.0 Tesla (3T) MRI.

Correction of Through-Plane Deformation Artifacts in Stimulated Echo Acquisition Mode Cardiac Imaging

Attempts to use a stimulated echo acquisition mode (STEAM) in cardiac imaging are impeded by imaging artifacts that result in signal attenuation and nulling of the cardiac tissue. In this work, we present a method to reduce this artifact by acquiring two sets of stimulated echo images with two different demodulations. The resulting two images are combined to recover the signal loss and weighted to compensate for possible deformation‐dependent intensity variation. Numerical simulations were used to validate the theory. Also, the proposed correction method was applied to in vivo imaging of normal volunteers (n = 6) and animal models with induced infarction (n = 3). The results show the ability of the method to recover the lost myocardial signal and generate artifact‐free black‐blood cardiac images. Magn Reson Med, 2006.

Ultrafast Localization of the Optic Disc Using Dimensionality Reduction of the Search Space

Optic Disc (OD) localization is an important pre-processing step that significantly simplifies subsequent segmentation of the OD and other retinal structures. Current OD localization techniques suffer from impractically-high computation times (few minutes/image). In this work, we present an ultrafast technique that requires less than a second to localize the OD. The technique is based on reducing the dimensionality of the search space by projecting the 2D image feature space onto two orthogonal (x- and y-) axes. This results in two 1D signals that can be used to determine the x- and y- coordinates of the OD. Image features such as retinal vessels orientation and the OD brightness and shape are used in the current method. Four publicly-available databases, including STARE and DRIVE, were used to evaluate the proposed technique. The OD was successfully located in 330 images out of 340 images (97%) with an average computation time of 0.65 seconds.

Active shape model with inter-profile modeling paradigm for cardiac right ventricle segmentation

In this work, a novel active shape model (ASM) paradigm is proposed to segment the right ventricle (RV) in cardiac magnetic resonance image sequences. The proposed paradigm includes modifications to two fundamental steps in the ASM algorithm. The first modification includes employing the 2D-principal component analysis (PCA) to capture the inter-profile relations among shapes neighboring landmarks and then model the inter-profile variations between the training set. The second modification is based on using a multi-stage searching algorithm to find the best profile match based on the best maintained profiles relations and thus the best shape fitting in an iterative manner. The developed methods are validated using a database of short axis cine bright blood MRI images for 30 subjects with total of 90 images. Our results show that the segmentation error can be reduced by about 0.4 mm and contour overlap increased by about 4% compared to the classical ASM technique with paired Students t-test indicates statistical significance to a high degree for our results. Furthermore, comparison with literature shows that the proposed method decreases the RV segmentation error significantly.

Strain-encoded MRI to evaluate normal left ventricular function and timing of contraction at 3.0 tesla

To define the reproducibility of strain‐encoded (SENC) magnetic resonance imaging (MRI) for assessment of regional left ventricular myocardial strain and timing of contraction in a 3T MRI system.

Myocardial segmentation using constrained multi-seeded region growing

Multi-slice short-axis acquisitions of the left ventricle are fundamental for estimating the volume and mass of the left ventricle in cardiac MRI scans. Manual segmentation of the myocardium in all time frames per each cross-section is a cumbersome task. Therefore, automatic myocardium segmentation methods are essential for cardiac functional analysis. Region growing has been proposed to segment the myocardium. Although the technique is simple and fast, non uniform intensity and low-contrast interfaces of the myocardium are major challenges of the technique that limit its use in myocardial segmentation. In this work, we propose a modified region growing technique that ensures reliable and fast myocardial segmentation of short-axis images. The proposed technique initializes the region growing process using different seed points. Then two types of spatial constraints are used to guarantee fast and accurate segmentation. The technique has been tested and validated quantitatively using a large number of images of different qualities. The results confirm the reliability and accuracy of the proposed technique.