University of Twente Student Theses

Login

Teaching a machine beauty : intelligent interactive evolution of abstract animations

Deenik, W. (2017) Teaching a machine beauty : intelligent interactive evolution of abstract animations.

[img]
Preview
PDF
18MB
Abstract:Generative art in the form of animations is used more and more in the current screen-filled world. If these animations would be generated with as much freedom as possible, some problems arise. More freedom means more ‘ugly’ animations are generated. Also, to steer the algorithm towards better animations, human feedback is needed, but giving feedback on animations takes time. A system is proposed and developed which can generate diverse abstract animations, while minimizing the number of unpleasant animations in its output, to (partly) solve the stated problems. This system uses a modified version of an algorithm developed by Karl Sims to render the animations, a genetic algorithm to improve those animations and a neural network acting as a filter to reject unpleasant animations from the system’s output. Feedback is gathered using a custom developed feedback interface. The system is evaluated and the performance of the filtering component determined in terms of precision, recall and F score in several different tests. Results showed that the system was successful in generating a relative high number of animations considered pleasant. It could also, to a certain extent, recognize and reject unpleasant animations in its output. Several suggestions are made for improving the system.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Human Media Interaction MSc (60030)
Link to this item:http://purl.utwente.nl/essays/73506
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page