Two automatic training-based forced calibration algorithms for left ventricle boundary estimation in cardiac images

Pixel classification algorithms based on temporal information, edge detection algorithms based on spatial information when used in combination are not sufficient for boundary estimation of the left ventricle (LV) in cardiovascular X-ray images. Poor contrast in the LV apex zone the fuzzy region in the inferior wall due to the overlap of the LV with the diaphragm, the inherent noise, and the variability of the modulation transfer function in X-ray imaging systems causes great difficulties in LV segmentation. To overcome the above problems, calibration algorithms were developed by Suri et al. (1996). These algorithms are training-based and provides a correction to the pixel-based classification or edge detection raw boundaries. This paper presents two training-based forced calibration algorithms for correcting the raw boundaries produced by classifiers. The authors force the raw LV contour to pass through the LV apex and then perform the calibration. Over a database of 377 patient studies having end-diastole and end-systole frames, the mean boundary error for the classifier system is 5.20 mm, the two forced calibration algorithms yield an error of 3.14 mm and 3.04 mm with a standard deviation of 2.73 mm and 2.89 mm.