Simulated annealing for motion correction in digital subtraction angiography

Simulated annealing, a heuristic technique for function optimization, is adapted to minimize the motion artifacts in digital subtraction angiography (DSA). The mask image is transformed before subtracting it from the contrast image, and the amount of misregistration between the transformed mask and the contrast is measured by a nonlinear function. This function is minimized in a discretized multidimensional space of one-to-one bilinear geometric transformations. Heuristic strategies are used to fix initial and subsequent temperatures, to decide on thermal equilibrium at a temperature, and to generate search points. The results are compared to those obtained by genetic algorithms.<>