Visualize and Paint GAN Activations

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Herdt, Rudolf, Maass, Peter
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929451621679104
author Herdt, Rudolf
Maass, Peter
author_facet Herdt, Rudolf
Maass, Peter
contents We investigate how generated structures of GANs correlate with their activations in hidden layers, with the purpose of better understanding the inner workings of those models and being able to paint structures with unconditionally trained GANs. This gives us more control over the generated images, allowing to generate them from a semantic segmentation map while not requiring such a segmentation in the training data. To this end we introduce the concept of tileable features, allowing us to identify activations that work well for painting.
format Preprint
id arxiv_https___arxiv_org_abs_2405_15636
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Visualize and Paint GAN Activations
Herdt, Rudolf
Maass, Peter
Computer Vision and Pattern Recognition
Machine Learning
We investigate how generated structures of GANs correlate with their activations in hidden layers, with the purpose of better understanding the inner workings of those models and being able to paint structures with unconditionally trained GANs. This gives us more control over the generated images, allowing to generate them from a semantic segmentation map while not requiring such a segmentation in the training data. To this end we introduce the concept of tileable features, allowing us to identify activations that work well for painting.
title Visualize and Paint GAN Activations
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2405.15636