Deep learning of individual aesthetics

Abstract

Accurate evaluation of human aesthetic preferences represents a major challenge for creative evolutionary and generative systems research. Prior work has tended to focus on feature measures of the artefact, such as symmetry, complexity and coherence. However, research models from Psychology suggest that human aesthetic experiences encapsulate factors beyond the artefact, making accurate computational models very difficult to design. The interactive genetic algorithm (IGA) circumvents the problem through human-in-the-loop, subjective evaluation of aesthetics, but is limited due to user fatigue and small population sizes. In this paper we look at how recent advances in deep learning can assist in automating personal aesthetic judgement. Using a leading artist's computer art dataset, we investigate the relationship between image measures, such as complexity, and human aesthetic evaluation. We use dimension reduction methods to visualise both genotype and phenotype space in order to support the exploration of new territory in a generative system. Convolutional Neural Networks trained on the artist's prior aesthetic evaluations are used to suggest new possibilities similar or between known high quality genotype-phenotype mappings. We integrate this classification and discovery system into a software tool for evolving complex generative art and design.

Complexity and Aesthetics in Generative and Evolutionary Art

“…Complexity measures, carefully chosen for specific styles or types of generative art can capture some broad aspects of personal aesthetic judgement, but they are insufficient alone to fully replace human judgement and discretion. Using other techniques, such as deep learning, may result in slightly better correlation to individual human judgement [33], however such systems require training on large datsets which can be tedious and time-consuming for the artist and still do not do as well as the trained artist's eye in resolving aesthetic decisions.…”

Interpretable Aesthetic Analysis Model for Intelligent Photography Guidance Systems

“…Since we usually have a limited number of training samples with human-labeled aesthetic scores, it is not practical to learn the feature extractor from scratch. Thus we start from a pre-trained ResNet [15] for its proven ability to learn robust image representations that show ever-improving performance in the task of aesthetic evaluation [17,36,37]. Feature maps extracted by the CNN are fed into attribute score estimators ,which output attribute scores.…”

Validation of an Aesthetic Assessment System for Commercial Tasks

“…The purpose of the computational aesthetic evaluation is to simulate the visual system and human perception to make an aesthetic judgment about the images automatically [ 10 ]. In recent years, many researchers from different fields of knowledge, such as Artificial Intelligence, Psychology, Arts or Design, have focused on the identification of the characteristics most related to human aesthetic preferences, as well as on the modelling of computer systems to recreate human evaluations for classification and prediction tasks [ 7 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 ].…”