Generating Kranok patterns with an interactive evolutionary algorithm.

The kranok pattern is a classical Thai pattern, often seen in the ornamentation of Thai religious artifacts such as Tripiṭaka cabinets, temple doors and coffins. All Thai trainee artists must practice drawing the kranok pattern because it is a fundamental motif in Thai traditional decorative art. Individual skilled artists and instructors have their own preferred ways of drawing the pattern, and their styles differ in ornamental details. Nevertheless, all make use of similar basic structures. Most trainees learn a range of styles by observing experienced artists. After exploring the methods of drawing used by experts and published in textbooks, we designed an algorithm for automatically generating kranok patterns. We used an interactive evolutionary algorithm (IEA) to improve the aesthetic appeal of the generated patterns in response to users' feedback. The aim of our work was not to replace artists with machines, but to enhance human artistic expression. Specifically, the work aimed to help users without artistic skills to create kranok patterns in their own style. The algorithm facilitated the creation of a variety of personalized kranok patterns – diverse in their expressive curvature, refinement and proportions – that were satisfying to the varying preferences of a range of users. We also analyzed the proposed algorithm's behavior in terms of its convergence to generate specific shapes. The proposed method was examined by 28 respondents (27 Thai and 1 foreign) who were selected to include representatives of both sexes as well as experts in both Thai drawing and evolutionary algorithms. The results from our questionnaires showed that all respondents were satisfied with the generated kranok patterns: one respondent was 'completely satisfied', seventeen 'very satisfied', seven 'moderately satisfied', and three 'slightly satisfied'. • Our IEA enables machines to create a variety of Kranok patterns. • The IEA takes each user' feedback to tailor the patterns specially for her. • We study the IEA's behavior and user satisfaction levels. • The IEA rapidly converges and generates desired patterns without annoying users. • From the surveys, most respondents are overall highly satisfied with the application. [ABSTRACT FROM AUTHOR]