In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Yiran, Shu, Zhixin, Smith, Cameron, Oh, Seoung Wug, Huang, Jia-Bin
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914751716524032
author Xu, Yiran
Shu, Zhixin
Smith, Cameron
Oh, Seoung Wug
Huang, Jia-Bin
author_facet Xu, Yiran
Shu, Zhixin
Smith, Cameron
Oh, Seoung Wug
Huang, Jia-Bin
contents 3D-aware GANs offer new capabilities for view synthesis while preserving the editing functionalities of their 2D counterparts. GAN inversion is a crucial step that seeks the latent code to reconstruct input images or videos, subsequently enabling diverse editing tasks through manipulation of this latent code. However, a model pre-trained on a particular dataset (e.g., FFHQ) often has difficulty reconstructing images with out-of-distribution (OOD) objects such as faces with heavy make-up or occluding objects. We address this issue by explicitly modeling OOD objects from the input in 3D-aware GANs. Our core idea is to represent the image using two individual neural radiance fields: one for the in-distribution content and the other for the out-of-distribution object. The final reconstruction is achieved by optimizing the composition of these two radiance fields with carefully designed regularization. We demonstrate that our explicit decomposition alleviates the inherent trade-off between reconstruction fidelity and editability. We evaluate reconstruction accuracy and editability of our method on challenging real face images and videos and showcase favorable results against other baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2302_04871
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing
Xu, Yiran
Shu, Zhixin
Smith, Cameron
Oh, Seoung Wug
Huang, Jia-Bin
Computer Vision and Pattern Recognition
3D-aware GANs offer new capabilities for view synthesis while preserving the editing functionalities of their 2D counterparts. GAN inversion is a crucial step that seeks the latent code to reconstruct input images or videos, subsequently enabling diverse editing tasks through manipulation of this latent code. However, a model pre-trained on a particular dataset (e.g., FFHQ) often has difficulty reconstructing images with out-of-distribution (OOD) objects such as faces with heavy make-up or occluding objects. We address this issue by explicitly modeling OOD objects from the input in 3D-aware GANs. Our core idea is to represent the image using two individual neural radiance fields: one for the in-distribution content and the other for the out-of-distribution object. The final reconstruction is achieved by optimizing the composition of these two radiance fields with carefully designed regularization. We demonstrate that our explicit decomposition alleviates the inherent trade-off between reconstruction fidelity and editability. We evaluate reconstruction accuracy and editability of our method on challenging real face images and videos and showcase favorable results against other baselines.
title In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2302.04871