HumanRAM: Feed-forward Human Reconstruction and Animation Model using Transformers

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Yu, Zhiyuan, Li, Zhe, Bao, Hujun, Yang, Can, Zhou, Xiaowei
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909635249700864
author Yu, Zhiyuan
Li, Zhe
Bao, Hujun
Yang, Can
Zhou, Xiaowei
author_facet Yu, Zhiyuan
Li, Zhe
Bao, Hujun
Yang, Can
Zhou, Xiaowei
contents 3D human reconstruction and animation are long-standing topics in computer graphics and vision. However, existing methods typically rely on sophisticated dense-view capture and/or time-consuming per-subject optimization procedures. To address these limitations, we propose HumanRAM, a novel feed-forward approach for generalizable human reconstruction and animation from monocular or sparse human images. Our approach integrates human reconstruction and animation into a unified framework by introducing explicit pose conditions, parameterized by a shared SMPL-X neural texture, into transformer-based large reconstruction models (LRM). Given monocular or sparse input images with associated camera parameters and SMPL-X poses, our model employs scalable transformers and a DPT-based decoder to synthesize realistic human renderings under novel viewpoints and novel poses. By leveraging the explicit pose conditions, our model simultaneously enables high-quality human reconstruction and high-fidelity pose-controlled animation. Experiments show that HumanRAM significantly surpasses previous methods in terms of reconstruction accuracy, animation fidelity, and generalization performance on real-world datasets. Video results are available at https://zju3dv.github.io/humanram/.
format Preprint
id arxiv_https___arxiv_org_abs_2506_03118
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle HumanRAM: Feed-forward Human Reconstruction and Animation Model using Transformers
Yu, Zhiyuan
Li, Zhe
Bao, Hujun
Yang, Can
Zhou, Xiaowei
Graphics
Computer Vision and Pattern Recognition
3D human reconstruction and animation are long-standing topics in computer graphics and vision. However, existing methods typically rely on sophisticated dense-view capture and/or time-consuming per-subject optimization procedures. To address these limitations, we propose HumanRAM, a novel feed-forward approach for generalizable human reconstruction and animation from monocular or sparse human images. Our approach integrates human reconstruction and animation into a unified framework by introducing explicit pose conditions, parameterized by a shared SMPL-X neural texture, into transformer-based large reconstruction models (LRM). Given monocular or sparse input images with associated camera parameters and SMPL-X poses, our model employs scalable transformers and a DPT-based decoder to synthesize realistic human renderings under novel viewpoints and novel poses. By leveraging the explicit pose conditions, our model simultaneously enables high-quality human reconstruction and high-fidelity pose-controlled animation. Experiments show that HumanRAM significantly surpasses previous methods in terms of reconstruction accuracy, animation fidelity, and generalization performance on real-world datasets. Video results are available at https://zju3dv.github.io/humanram/.
title HumanRAM: Feed-forward Human Reconstruction and Animation Model using Transformers
topic Graphics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.03118