Optimization of the global matching between two contours defined by ordered points using an algorithm based on dynamic programming

Abstract

This paper presents a new assignment algorithm with order restriction, developed using the paradigm of dynamic programming. The algorithm was implemented and tested to determine the best global matching between two sets of points that represent the contours to be matched. In the experimental tests done, we used the affinity matrix obtained via the method proposed by Shapiro based on geometric modeling and modal matching.
The proposed algorithm revealed an optimum performance, when compared with the classic assignment algorithms considered in this work: Hungarian method, Simplex for Flow Problems and LAPm. Indeed, the quality of the matching improved when compared with these three algorithms, because the crossed matching, allowed by the conventional assignment algorithms, disappeared. Besides, the computational cost of our algorithm is very low in comparison with the other three, resulting in lesser execution times.