Temporal and spatial registration for cardiac strain rate imaging

Current strain rate imaging (SRI) suffers from Doppler principle limitations. Here we present an alternative 2D correlation-based speckle tracking method for SRI. To fully characterize local deformation over a cardiac cycle with Lagrangian strain estimates, in addition to spatial registration by referencing displacement measurements back to its original geometry, we perform temporal registration to compensate for different sampling times between beams. This algorithm was tested on both simulations and measurements from a porcine cardiac model. Simulations show strain rate errors of 30% of maximum without time registration. Strain rate with temporal and spatial registration show good agreement with theoretical predictions. In addition, the tracking algorithm is robust, producing almost no residual error in displacements accumulated over one complete cardiac cycle. Preliminary results from a porcine interventricular septum also show high repeatability in both longitudinal displacement and strain rate estimates between cardiac cycles.