A Feasibility Study of 3D Motion Compensation in 3D Diverging Wave Compounding

Abstract

Matrix-array-based 3D ultrasound imaging is intensively studied in research communities. However, 3D ultrasound imaging still suffers from motion artifacts. Eliminating motion artifacts is one of the keys to improve image quality for 3D ultrasound imaging. Current motion compensation (MoCo) methods using Doppler estimation only allow to compensate displacements in the beamformed lines, i.e., 1D MoCo. In this work, we propose 3D MoCo that estimates the 3D velocity fields for 3D diverging wave compounding imaging using a round-trip scan sequence. The framework employs two constraints, i.e., Doppler and optical flow in a round-trip scan sequence, and two regularization terms, i.e., the free divergence and the first order smoothness of the velocity fields, to formulate the estimation of the 3D velocity fields as a global optimization problem. The in-silico experiments were conducted by establishing and moving a cystic cuboid phantom in a prefixed direction at given velocities. Both the results of low and high velocity amplitudes demonstrate the feasibility of the proposed 3D MoCo method. Our 3D MoCo method can achieve higher image contrast and preserve speckle patterns better than the 1D MoCo method, showing a promising potential for high quality 3D ultrasound imaging.