Measuring the impact of subjective and transient aesthetics in the generation and appreciation of 3D virtual artwork

Abstract

The topic of aesthetics is a vast puzzle with many contributing parts. It often seems
incomprehensible; however, through decomposition into smaller problems, a solution may
be within reach. It impacts almost every aspect of life, including romantic and platonic
relationships, architecture, decoration, writing, theatre, mathematical proofs, algorithms,
product design and artwork.

This thesis aims to partially deconstruct the individual aesthetic judgement of sculptures
into its component parts through the steps of understanding these parts at both individual
and collective levels, introducing formalised measures which calculate how well a specific
sculpture exhibits a particular aesthetic feature and then using these measures to provide
a novel, extendable and flexible model of individual aesthetic judgement.

This project represents a step forward in the computational research of aesthetics and
aesthetic judgement through original Virtual Reality Environments and an art generation
process capable of creating interesting 2D and 3D artwork. Experiments reveal some key
aspects that contribute to human aesthetic judgement, both positively and negatively. The
experiments also resulted in clear rules to follow for the generation of sculptures according
to personal preference.

New aspects are formalised, calculating how connected, calm, or friendly a sculpture
appears; all three are found to be closely aligned with the human understanding of the
term. Finally, a theoretical model of individual aesthetic preference is presented. The
model is based on three categories of aesthetic aspects: constant, subjective, and transient;
these allow the generation of sculptures which are unique to an individual. Importantly, the
models can be applied to replicating a specified style of artwork, meaning that new items
can be generated that exhibit close relationships to the original style whilst still being novel.
Constituting a multi-pronged approach, the contributions presented in the thesis move us
closer to decoding the puzzle of aesthetic appreciation.

Date of Award	Sept 2023
Original language	English
Awarding Institution	Aston University
Supervisor	Aniko Ekárt (Supervisor) & Ulysses Bernardet (Supervisor)