Understanding Aesthetics and Fitness Measures in Evolutionary Art Systems

Abstract

One of the general aims of evolutionary art research is to build a computer system capable of creating interesting, beautiful, or creative results, including images, videos, animations, text, and performances. In this context, it is crucial to understand how fitness is conceived and implemented to explore the “interestingness,” beauty, or creativity that the system is capable of. In this paper, we survey the recent research on fitness for evolutionary art related to aesthetics. We also cover research in the psychology of aesthetics, including relation between complexity and aesthetics, measures of complexity, and complexity predictors. We try to establish connections between human perception and understanding of aesthetics with current evolutionary techniques.

The Enigma of Complexity

“…The application of complexity measures and their relation to aesthetics in generative and evolutionary art are numerous (see e.g. [14] for an overview). A number of researchers have tested complexity measures as candidates for fitness measures in evolutionary art systems for example.…”

“…More recent surveys and analysis of computational aesthetics trace the history [9,12] and current state of research in this area [14]. Other approaches introduce features such as symmetry as a counterbalance to complexity, situating aesthetic appeal somewhere within the range spanning between these two properties [30].…”

Deep learning of individual aesthetics

“…We first applied image analysis methods to the 2D images, looking for correlations between aesthetic rating or categorisation and overall image properties. The aim of this analysis was to better understand if any common factors, such as symmetry, complexity or morphological structure were influential on personal aesthetic decisions, as these factors are commonly used in general aesthetic measures [18]. The results of this analysis is presented in Section 3.3.…”

“…However, the majority of evolutionary algorithms used for art and design rely on being able to compute a fitness measure of phenotype aesthetics. Devising formal aesthetic measures is a longstanding, but generally illusive quest in evolutionary computing and psychology research [5,16,18].…”

“…Over the years, the research community has proposed many new theories and measures of aesthetics, with research from both the computational aesthetics (CA) and psychology communities [18]. Despite much effort and many advances, a computable, universal aesthetic measure remains an open problem in evolutionary music and art research [31].…”