Motion estimation using prebeamformed ultrasound signals

In many cases, motion information about tissue or organs carries relevant information to complete the diagnosis based on ultrasound imaging (e.g. elastography, CFM, heart motion analysis etc.). Often, the displacement has a very small amplitude, for example when estimating motion of slow moving flow or when the frame rate is very high. In this study, we propose a method adapted to very small displacement estimation, about 1/10 of the wavelength. This method is inspired by multi-beam methods. It uses multiple ID displacements along prebeamformed signals, which are then recombined to estimate the 2D displacement. Other multi-beam techniques are based on a small number of ID estimations. Using more ID estimations strongly improves the precision of the estimated displacement field. The estimation is done in two steps: 1D displacements are estimated along prebeamformed signals focused around the point of interest. 1D displacements are then recombined to form the 2D displacement using a motion model. We show that 1D displacements are projections of the real displacement of the point. The method is validated both experimentally on a large set of translated phantom. In both simulation and experimental cases, we show that our method leads to an average 1μm × 100nm mean error along transverse and axial direction, compared to 5μm × 1μm along transverse and axial direction for classical speckle-tracking method.