The technology of image matching by the criterion of conformity of image fragments samples
Vladimir Alekseyevich Fursov, A V Gavrilov, Yegor V. Goshin, Kirill Pugachev · Journal of Physics Conference Series · 2018
In this paper we considered the image matching problem. The problem is solved in the framework of an area approach based on the comparison of image fragments in the neighborhood of the points being compared. We proposed a new criterion of image fragments similarity based on the principle of image fragments samples conformity. An important advantage of the method is the possibility of comparison of very small image fragments. However, computation of this criterion is very time-consuming. So, we proposed parallel algorithm of conforming image matching using MPI. We evaluated acceleration and efficiency of parallel algorithm. The results of experiments on the "Cones" test images show high accuracy of the method.