Knowledge-guided automatic segmentation of the left ventricle from MR

Abstract

The routinely used clinical practice of manual tracing of the blood pool from short axis cine MR images to compute ejection fraction (EF) is cumbersome, time consuming, and operator dependent. In this paper we present an algorithm that automatically segments the left ventricle (LV) using the a priori knowledge of the intensity responses of the tissue in different MR modalities, along with the LV morphology. Our method for the automatic computation of the EF is based on segmenting the left ventricle by combining the fuzzy connectedness and the physics-based deformable model frameworks. We have validated our method against manual delineation performed by experienced radiologists on the data from nine asymptomatic volunteers with very encouraging results.