Supervised Genetic Search for Parameter Selection in Painterly Rendering

Abstract

Abstract. This paper investigates the feasibility of evolutionary search techniques as a mechanism for interactively exploring the design space of 2D painterly renderings. Although a growing body of painterly rendering literature exists, the large number of low-level configurable parameters that feature in contemporary algorithms can be counter-intuitive for non-expert users to set. In this paper we first describe a multi-resolution painting algorithm capable of transforming photographs into paintings at interactive speeds. We then present a supervised evolutionary search process in which the user scores paintings on their aesthetics to guide the specification of their desired painterly rendering. Using our system, nonexpert users are able to produce their desired aesthetic in approximately 20 mouse clicks -around half an order of magnitude faster than manual specification of individual rendering parameters by trial and error.

Eugene: A Generic Interactive Genetic Algorithm Controller

“…For example, should they be tools for those who wish to explore a solution space, or should they be used when there are clearly defined user goal states? [11] Issues such as the fitness evaluation bottleneck are discussed by [12] who also considers the impact of fatigue.…”

Evolving assemblages

“…The use of evolutionary algorithms to create image filters and non-photorealistic renderings of source images has been explored by several researchers. Focusing on the works where there was an artistic goal, we can mention the research of: Ross et al (2006) and Neufeld et al (2007), where genetic programming (GP) (Koza, 1992), multi-objective optimisation techniques, and an empirical model of aesthetics are used to automatically evolve image filters; Lewis (2004), which evolved live-video processing filters through interactive evolution; Machado et al (2002), where GP is used to evolve image colouring filters from a set of examples; Yip (2004), which employs GAs to evolve filters that produce images that match certain features of a target image; Collomosse and Hall (2005) and Collomosse (2006Collomosse ( , 2007, which use image salience metrics to determine the level of detail for portions of the image, and GAs to search for painterly renderings that match the desired salience maps; Hewgill and Ross (2003) use GP to evolve procedural textures for 3D objects.…”

Stroke Based Painterly Rendering

“…A mapping was established between this space, and a higher dimensional parameter space of an impasto oil painterly algorithm [21] (Figure 1.9). Another option is to let the painterly rendering system provides some potential paintings and let the user rate them [2]. Then new paintings are generated taking into account the score of each candidates, using an interactive evolutionary algorithm.…”