[4Xin2] Poster session 2

[4Xin2-03] Generation of Abstract Painting Like Textures Using Ant Painting by Iterative Evolutionary Computation

〇Koichiro Sato¹, Kazunori Mizuno² (1.Graduate School, Takushoku University, 2.Takushoku University)

Keywords:Image Editing, Evolutionary Computation, Multi Agent System

Abstract paintings are works of art that are created by combining dots, lines, and figures. Generally, it is common to generate abstract paintings using image editing software, but it is an effort for those who do not have editing skills to process the image to their preference. In addition, conventional methods of creating abstract images often require a limited number of operations, which limits the range of expression and can become monotonous. Therefore, we propose a method for processing input images into a variety of abstract-style textures. Specifically, we propose a texture generation method called "Ant Painting," which uses the emergent phenomenon of artificial ants to create abstract painting-style textures while preserving the characteristics of the input image. The parameters of the artificial ants are tuned according to an iterative evolutionary computation. We provide an example of execution using an input image to demonstrate that the proposed method is effective in rendering a variety of abstract textures while keeping the user's preferences.