Parametric estimation of 2-D motion on ultrasonic images using respiratory signal

It is usually difficult to estimate two-dimensional (2-D) motion from ultrasonic images because of their low image quality. In this paper, we propose a motion estimation method in which respiratory information strongly correlated with the motion of an internal organ is used. We represent the 2-D motion as a regressive model in which, with respect to the respiratory signal, we introduce an appropriate temporal constraint. Additionally, for this regressive model we construct an algorithm to determine the parameters in the model as a maximum likelihood (ML) estimation and to determine the motion as a MAP estimation.