Automatic detection of echocardiographic LV-contours-a new image enhancement and sequential tracking method

摘要

The automatic analysis of 2-D echocardiographic images in order to evaluate the cardiac function quantitatively, detect wall-motion abnormalities, and perform 3-D modeling and simulation of the left ventricle (LV) would constitute a valuable tool for clinical use. All these analyses require the complete and reliable determination of at least the LV endocardial border. The authors describe a method for accurate detection of this LV border in a series of short- and long-axis views in stop images at end-systole and end-diastole, using low-computational-cost algorithms. The method features a powerful enhancement procedure and a contour-tracking algorithm using a secondary search process, where needed. Tests on low-quality images indicate good performance.