A fast strategy for image matching using Hausdorff distance

In this paper, the problem of matching the modal in an image with a fast search strategy is discussed. While the Hausdorff measure yields a method to comparing edge images that is robust to object occlusion, image noise, and clutter, etc, it requires high computational complexity in computing the Hausdorff distance between two edge point sets. A fast search strategy is proposed to reduce the computational load in this paper. A similarity measure is defined in the proposed strategy. Experiments are performed on a large test set and the results show that the fast search strategy diminishes the number of positions that are in need of calculating Hausdorff distance to about one sixth of that in origination. Thus the computational load is reduced and it is helpful to use Hausdorff distance in real time image matching.