Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Xiao, Junjin, Zhang, Qing, Xu, Zhan, Zheng, Wei-Shi
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2403.10335
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929278213423104
author Xiao, Junjin
Zhang, Qing
Xu, Zhan
Zheng, Wei-Shi
author_facet Xiao, Junjin
Zhang, Qing
Xu, Zhan
Zheng, Wei-Shi
contents Human avatar has become a novel type of 3D asset with various applications. Ideally, a human avatar should be fully customizable to accommodate different settings and environments. In this work, we introduce NECA, an approach capable of learning versatile human representation from monocular or sparse-view videos, enabling granular customization across aspects such as pose, shadow, shape, lighting and texture. The core of our approach is to represent humans in complementary dual spaces and predict disentangled neural fields of geometry, albedo, shadow, as well as an external lighting, from which we are able to derive realistic rendering with high-frequency details via volumetric rendering. Extensive experiments demonstrate the advantage of our method over the state-of-the-art methods in photorealistic rendering, as well as various editing tasks such as novel pose synthesis and relighting. The code is available at https://github.com/iSEE-Laboratory/NECA.
format Preprint
id arxiv_https___arxiv_org_abs_2403_10335
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle NECA: Neural Customizable Human Avatar
Xiao, Junjin
Zhang, Qing
Xu, Zhan
Zheng, Wei-Shi
Computer Vision and Pattern Recognition
Human avatar has become a novel type of 3D asset with various applications. Ideally, a human avatar should be fully customizable to accommodate different settings and environments. In this work, we introduce NECA, an approach capable of learning versatile human representation from monocular or sparse-view videos, enabling granular customization across aspects such as pose, shadow, shape, lighting and texture. The core of our approach is to represent humans in complementary dual spaces and predict disentangled neural fields of geometry, albedo, shadow, as well as an external lighting, from which we are able to derive realistic rendering with high-frequency details via volumetric rendering. Extensive experiments demonstrate the advantage of our method over the state-of-the-art methods in photorealistic rendering, as well as various editing tasks such as novel pose synthesis and relighting. The code is available at https://github.com/iSEE-Laboratory/NECA.
title NECA: Neural Customizable Human Avatar
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.10335