The application of simulated annealing to a map seeking circuit

This paper examines how to apply simulated annealing to a map seeking circuit (MSC) when the circuit is formulated as a gradient ascent optimization problem. This modification is performed on a map seeking circuit solving the three dimensional model correspondence problem. It is shown that when the input image is partially occluded and covered by random edges, then the MSC can converge incorrectly to a local maximum instead of the global maximum. Under these conditions, simulated annealing is shown to improve the performance of the MSC through a series of experiments involving the Hausdorff distance metric.