Cardiac Motion Estimation from 3D Echocardiography with Spatiotemporal Regularization

Cardiac deformation and motion analysis is important for studying heart function and mechanics. Deformation and motion abnormality of the myocardial wall is usually associated with ischemia and infarct. Three-dimensional (3D) echocardiographic (echo) imaging is themost widely used method to estimate cardiac motion. However, quantitative motion analysis from echo images is still a challenging problem due to the complexity of cardiac motion, limitations in spatial and temporal resolutions, low signal noise ratio and imaging artifacts such as signal dropout. We developed a novel method to quantitatively analyze cardiac deformation andmotion from echo sequences. Our estimated cardiac motion is not only regularized to be spatially but also temporally smooth. We validate our methods using (1) simulated echo images with known ground truth, and (2) in vivo echo images acquired on open-chests pigs with sonomicrometry. Tests indicate that our method can estimate cardiac motion more accurately than methods without temporal regularization.