When a specialist scans your heart motion with an ultrasonic probe, usually they have to use subjective analysis to determine any heart problems. Angina, for example, is when one part of the heart wall expands more than the rest. A trained specialist can detect this by watching the scan, but there was no way to quantify the amount of distortion, because the heart moves all over the place when its beating.

My software would attempt to match up significant points in sequential scans to detect the absolute motion of that point over time. It did this by comparing the speculated motion with the motion of other nearby points, and attempting to find a set of correlations that resulted in smooth motion for all points. Once it had the absolute path each point took, it could plot the lengths of those paths against their position along the heart wall to provide a direct measure of heart wall motion abnormality.

This software, like all I wrote at Yale, was in Pascal on a VAX Microstation running VMS. I was also the administrator for that system.