Sole Arch Width Extraction Using Greedy Algorithm

In footwear production, the size of sole should be checked to match the corresponding upper before combining them. At present, manual recognition method and traditional machine vision method is not able to recognize unlabelled sole. In this paper, we propose a novel method to extract the feature parameter of unlabelled sole to recognize its size. The feature parameter is the arch width of the sole which has a quality with shape invariance. An innovative greedy algorithm was proposed to get the arch width. The algorithm adopts successive approximation method to extracted the shortest width of sole. Experimental results show that the arch width extraction method proposed in this paper can be effectively used to recognize the size of sole.