Obtaining blood velocity profile from coronary arteriograms via optimally controlled optical flow

Physical constraints are used to adapt known optical flow algorithms to study blood circulation in healthy and stenosed arteries by tracking dye dispersion using radiological image sequences. To this end, the multiplier methods of constrained optimization are put forward. Instabilities are handled through the Lions regularization process. Simulated radiological images taking into account blood flow and contrast substance dispersion are computed to test the proposed algorithms. Results from simulations as well as from clinical radiological data are presented. Results on simulated images indicate a good fluid flow estimation. Results on clinical images appear to be qualitatively correct. The proposed method has many advantages. It allows the introduction of any physical a priori knowledge modeled with partial differential equations, permits clear distinction between mathematical (smoothness) and physical constraints, and provides efficient image flow algorithms.<
>