A pattern recognition approach to automated coronary calcium scoring

An automated method for coronary calcification detection is presented. First the heart region is extracted, in which objects potentially representing calcifications are obtained by thresholding. Besides coronary calcifications, the set of objects includes other heart calcifications, bony structures and noise. For each object, features describing its size, shape, position and appearance are computed. Several classifiers and classification strategies are evaluated. Best results are obtained with a specifically designed sequence of kNN classifiers that employ sequential forward feature selection. First obvious non-calcifications are removed, then calcifications are distinguished from non-calcifications and a final classifier discerns coronary calcifications from other cardiac calcifications. In 14 CT scans containing 61 coronary calcifications, 46 (75%) are detected at the expense of on average 0.9 false positive objects per scan.