Adaptive speckle tracking algorithms for improved ultrasound blood flow imaging
Abstract
This thesis will present several ways to further improve the robustness of the speckle tracking method. The speckle tracking algorithm was made more adaptive to the velocity changes over time, utilizing velocity estimates from the previous flow frame to update the placement and size of the search region. The size of the search region was adjusted to the velocity change from the two previous flow frames, giving a larger search region for a higher acceleration. The tracking grid was also made finer according to the size of the search region, where the distance between the interpolated tracking points was made smaller. Multi-lag tracking was also tested, utilizing previous velocity estimates to adaptively find the best tracking lag over time. A higher tracking lag for small velocities should increase the accuracy of the estimated velocities. The different methods to improve the robustness of the speckle tracking algorithm was tested with the use of simulated data of a carotid artery from the ultrasound simulation code, Field II. The data collected from the ultrasound simulation code is based on plane wave imaging and parallel receive beam forming to achieve a high frame rate. Results show that the methods tested will improve the tracking estimates, where multi-lag tracking improves the tracking estimates for small velocities.