Analysis of left ventricular motion using a general robust point matching algorithm

Abstract

In this paper we present a novel and fully automated approach for the estimation of non-rigid cardiac motion and deformation from sequences of three dimensional medical images. Our approach extends the robust point matching (RPM) algorithm to include shape-based information as inspired by our own previous work in this area. The resulting algorithm labeled as generalized robust point matching (G-RPM) is capable of accurately estimating left ventricular motion without the need of a prior, and often time-consuming, segmentation of the myocardium. We evaluate our approach on both synthetic data as well as using sequences of in-vivo cardiac magnetic resonance images. The approach can easily be adjusted for a number of applications to find the optimal non-rigid transformation. © Springer-Verlag Berlin Heidelberg 2003.