Estimation of Motion Parameters for Ultrasound Images Using Motion Blur Invariants

Abstract: The quality of fetal ultrasound images is significantly affected by motion blur while the imaging system requires low motion quality in order to capture accurate data. This can be achieved with a mathematical model of motion blur in time or frequency domain. We propose a new model of linear motion blur in both frequency and moment domain to analyse the invariant features of blur convolution for ultrasound images. Moreover, the model also helps to provide an estimation of motion parameters for blur length and angle. These outcomes might imply great potential of this invariant method in ultrasound imaging application.