Image approximation in a limited class of graphic primitives

The paper is devoted to the development and study of a unified approach to approximating raster images with sets of graphic primitives of one class using multidimensional continuous optimization methods. The problem of image approximation is posed as a problem of finding an optimal set of primitive parameters that maximizes the objective function based on one of the metrics used to determine the similarity of two given images. Practical implementation based on the proposed approach is a modular software system in the Python programming language with graphical and command interfaces, functions for drawing, saving/loading results, and generating animated images. During the experiments, the efficiency of the algorithms used and the influence of their parameters and system settings on the quality of approximation were studied.