3D Generative Model Latent Disentanglement via Local Eigenprojection

Simple item page
dc.contributor.authorFoti, Simoneen_US
dc.contributor.authorKoo, Bongjinen_US
dc.contributor.authorStoyanov, Danailen_US
dc.contributor.authorClarkson, Matthew J.en_US
dc.contributor.editorHauser, Helwig and Alliez, Pierreen_US
dc.date.accessioned2023-10-06T11:58:49Z
dc.date.available2023-10-06T11:58:49Z
dc.date.issued2023
dc.description.abstractDesigning realistic digital humans is extremely complex. Most data‐driven generative models used to simplify the creation of their underlying geometric shape do not offer control over the generation of local shape attributes. In this paper, we overcome this limitation by introducing a novel loss function grounded in spectral geometry and applicable to different neural‐network‐based generative models of 3D head and body meshes. Encouraging the latent variables of mesh variational autoencoders (VAEs) or generative adversarial networks (GANs) to follow the local eigenprojections of identity attributes, we improve latent disentanglement and properly decouple the attribute creation. Experimental results show that our local eigenprojection disentangled (LED) models not only offer improved disentanglement with respect to the state‐of‐the‐art, but also maintain good generation capabilities with training times comparable to the vanilla implementations of the models. Our code and pre‐trained models are available at .en_US
dc.description.number6
dc.description.sectionheadersORIGINAL ARTICLES
dc.description.seriesinformationComputer Graphics Forum
dc.description.volume42
dc.identifier.doi10.1111/cgf.14793
dc.identifier.issn1467-8659
dc.identifier.urihttps://doi.org/10.1111/cgf.14793
dc.identifier.urihttps://diglib.eg.org:443/handle/10.1111/cgf14793
dc.publisher© 2023 Eurographics ‐ The European Association for Computer Graphics and John Wiley & Sons Ltd.en_US
dc.rightsAttribution 4.0 International License
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.subjectdisentanglement
dc.subjectgenerative adversarial networks
dc.subjectgeometric deep learning
dc.subjectvariational autoencoder
dc.title3D Generative Model Latent Disentanglement via Local Eigenprojectionen_US
Files
Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
v42i6_08_14793.pdf
Size:
1.47 MB
Format:
Adobe Portable Document Format
Download
Collections
42-Issue 6