Generation Method of the Most Suitable Template Shape in the Template Matching: Study Using Genetic Algorithm that Takes Diversity into Consideration

Abstract

Template matching is an effective method for calculating the similarity of the corresponding positions between images. However, a mismatch may occur due to occlusions, shadows, or distortions. This paper shows that there is an optimal template shape based on the residual probability distribution function obtained from the searched image and proposes a method for automatically creating the optimal template shape. Moreover, this paper also presents the results of a study on the techniques used to maintain the diversity of each individual body when applying a genetic algorithm on a template. In order to evaluate the effectiveness of the proposed method, 5 different types of images are used for comparison with other three methods. In the experiments, the proposed method gave the best results compared to the other 3 methods, showing its high level of usefulness.