Phase Aberration Estimation in Synthetic Transmit Aperture Ultrasound Imaging and Its Application to Estimating Sound Speed

Despite the broad application of ultrasound imaging in modern diagnostic modalities, it often suffers from suboptimal image quality. Phase aberration is one of the main contributors to image degradation, and it appears as poor contrast, poor lateral resolution, and fill-in into hypoechoic regions due to the degraded beam-focusing quality. Image reconstruction is usually performed under the assumption of a homogeneous medium. Nonetheless, in the presence of spatial sound-speed heterogeneity, this hypothesis is no longer valid and leads to errors in the estimated echo arrival time. This dissertation investigates phase aberration estimation methods in synthetic transmit aperture ultrasound imaging (STA) and its application in estimating the speed map of the medium. STA Radio-frequency (RF) data were simulated and also acquired in experiments. However, the signal-to-noise ratio (SNR) of STA signals was much lower than that in B-mode. Therefore, we first developed a Filtered-Normalized-Cross-Correlation (F-NCC) method to estimate the phase aberration in noisy STA data. A 2D filter was applied in the temporal and spatial frequency domain to reduce the noise and off-axis signals, and its performance was validated with both simulation and experiment data. Then we derived an equation to relate the phase aberration to the average speed at a point in the medium. This equation was applied to estimate the average sound speed in the medium. It was demonstrated with a two-layered phantom that the average speed map can be used to estimate the local speed in layered objects, such as in the presence of subcutaneous fat and connective tissues. The image reconstructed based on the speed map across the medium had an improved quality globally. Furthermore, the local speed could be utilized as a biomarker. Phase aberration estimation and speed map methods were more successful when they were iterated. In the above studies, the focal point was selected in the first iteration, and it was not updated later. An adaptive localization method in which the focal point was updated by choosing the maximum brightness point within the selected area in each iteration was proposed. The results of both the simulation and experimental studies showed that adaptive localization improves the phase aberration estimation by 80% and the average speed estimation by 60%. In the future work, the potential to apply the phase aberration estimation method to estimate the local speed in non-layered objects was discussed.