We are developing algorithms to automatically (i.e., without human intervention) detect the borders of structures in cardiovascular images obtained from ultra-fast CT, 2-D ultrasound, intra-vascular ultrasound, or MR. The goal is to automatically detect the cardiac borders, or the vessel walls and structures, in disease and health. From these, three-dimensional models of the cardiovascular structures are derived, and quantitative evaluations of the mechanics and flow waveforms are derived.