Variational techniques for cardiac image analysis: Algorithms and applications.

摘要

In this dissertation we investigate several image segmentation and registration techniques based on the variational formulation for medical imaging applications. The five main research results are summarized below.First, a novel segmentation approach is proposed to jointly delineate the boundaries of epiand endocardium of the left ventricle on Magnetic Resonance Imaging (MRI) under a variational framework using level sets. While most left ventricle segmentation approaches incorporate a shape prior obtained by a training process from an ensemble of examples, we exploit a novel shape constraint using an implicit shape prior knowledge, which assumes shape similarity between epi- and endocardium allowing a variation under the Gaussian distribution.Second, we examine multi-modal data integration with an electroanatomic mapping (EAM) data and MRI images for computer-aided catheter ablation of atrial fibrillation accurately. Specifically, we propose a variational formulation for surface reconstruction and incorporate the prior shape knowledge, which results in a level set method. The proposed method enables simultaneous reconstruction and registration under nonlinear deformation.Third, A fully automated registration method is presented utilizing geometric features from a reliable segmentation of gated myocardial perfusion SPECT (MPS) volumes, where regions of myocardium and blood pools are extracted and used as an anatomical mask to de-emphasize the inhomogeneities of intensity distribution caused by perfusion defects and physiological variations. A multi-resolution approach is employed to represent coarse-to-fine details of both volumes. The extracted voxels from each level are aligned using a similarity measure with a piecewise constant image model and minimized using a gradient descent method. We then perform limited nonlinear registration of gated MPS to adjust for phase differences by automatic cardiac phase matching between CT and MPS. For phase matching, we incorporate nonlinear registration using thin-plate-spline-based warping.Fourth, a nonlinear ultrasound image registration method is proposed using the intensity and the local phase information under a variational framework. One application of this technique is to register two consecutive images in an ultrasound image sequence. Although intensity is the most widely used feature in traditional ultrasound image registration algorithms, speckle noise and lower image resolution make the registration process difficult. By integrating the intensity and the local phase information, we can find and track the nonlinear transformation of each pixel under diffeomorphism between the source and target images.Finally, we develop a fully automatic and accurate nonlinear volume registration for longitudinal Coronary CT angiography (CCTA) scan pairs. The proposed algorithm combines global displacement and local deformation using nonlinear volume co-registration with a volume-preserving constraint. Histogram matching of intensities between two serial scans is performed prior to nonlinear co-registration with dense nonparametric diffeomorphism in which sum of squared difference is used as a similarity measure. The segmented coronary artery trees provide initial anatomical landmarks for the co-registration algorithm that help localize and emphasize the structure of interest. To avoid possible bias caused by incorrect segmentation, we convolve the Gaussian kernel with the segmented binary coronary tree mask and define an extended weighted region of interest. A multi-resolution approach is employed to represent coarse-to-fine details of both volumes. The energy functional is optimized using a gradient descent method.Extensive computer simulations have been conducted and clinical validations have been performed to demonstrated the improved accuracy of the proposed techniques.