Edge Matching Based on Chamfer Distance Transform and Label Map

A novel method-label mapping is proposed to deal with the problem of noise in 3-4chamfer distance map.Features of edge pixels,which are difficult to be applied directly with the distance map,are available by label map.At the same time,the influence of noise in distance map is weakened due to the mapping between labels and pixels.Model-based method is used in edge matching and the matching score is evaluated by the distance map and label map.This process consists in searching for the best parameters of affine transform.A genetic algorithm is used in the estimation of these parameters.