Approximate license plate string matching for vehicle re-identification

From the empirical studies, it is quite difficult for the license plate recognition to perform 100% accuracy in a real world environment. Nevertheless, it is common that only a few characters are misread from a license plate recognition system. In this paper, license plate matching is used for vehicle re-identification. We evaluate several approximate string matching techniques to determine an applicable method for the domain of license plate matching. We also propose a weighted approximate string matching to further improve the performance. Our experimental results show that the generalized edit distance (GED) yields highest precision while the 2-Grams yields the highest recall. By weighting the distance cost from those two methods, our proposed weighted method gives the best performance.