Controlled Modification of Generated (Style)GAN Latent Vectors

Bachelor thesis (2019)

Authors

Y. Liu Electrical Engineering, Mathematics and Computer Science

Contributors

T.J. Viering Pattern Recognition and Bioinformatics - EEMCS (coach)
D.M.J. Tax Pattern Recognition and Bioinformatics - EEMCS (mentor)
Faculty
Electrical Engineering, Mathematics and Computer Science, Electrical Engineering, Mathematics and Computer Science
Copyright: © 2019 Yifei Liu
More Info
expand_more

Published Date

28-06-2019

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Electrical Engineering, Mathematics and Computer Science

Abstract

StyleGAN is a neural network architecture that is able to generate photo-realistic images. The diversity of generated images are ensured by latent vectors. These latent vectors encodes important features of generated images. They provide us insight-full information about properties of image generation in StyleGAN, which may also occur similarly in other neural network architectures. Using pre-trained StyleGAN models, we have conducted several experiments to show properties on StyleGAN based on results of style modification. The experiments have shown the influence of noise and styles in StyleGAN and how latent vectors can be manipulated in generated images.

Files

BEP_paper_1_.pdf
(.pdf | 12.6 Mb)