DreamComposer: Controllable 3D Object Generation via Multi-View Conditions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yang, Yunhan, Huang, Yukun, Wu, Xiaoyang, Guo, Yuan-Chen, Zhang, Song-Hai, Zhao, Hengshuang, He, Tong, Liu, Xihui
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909150626185216
author Yang, Yunhan
Huang, Yukun
Wu, Xiaoyang
Guo, Yuan-Chen
Zhang, Song-Hai
Zhao, Hengshuang
He, Tong
Liu, Xihui
author_facet Yang, Yunhan
Huang, Yukun
Wu, Xiaoyang
Guo, Yuan-Chen
Zhang, Song-Hai
Zhao, Hengshuang
He, Tong
Liu, Xihui
contents Utilizing pre-trained 2D large-scale generative models, recent works are capable of generating high-quality novel views from a single in-the-wild image. However, due to the lack of information from multiple views, these works encounter difficulties in generating controllable novel views. In this paper, we present DreamComposer, a flexible and scalable framework that can enhance existing view-aware diffusion models by injecting multi-view conditions. Specifically, DreamComposer first uses a view-aware 3D lifting module to obtain 3D representations of an object from multiple views. Then, it renders the latent features of the target view from 3D representations with the multi-view feature fusion module. Finally the target view features extracted from multi-view inputs are injected into a pre-trained diffusion model. Experiments show that DreamComposer is compatible with state-of-the-art diffusion models for zero-shot novel view synthesis, further enhancing them to generate high-fidelity novel view images with multi-view conditions, ready for controllable 3D object reconstruction and various other applications.
format Preprint
id arxiv_https___arxiv_org_abs_2312_03611
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle DreamComposer: Controllable 3D Object Generation via Multi-View Conditions
Yang, Yunhan
Huang, Yukun
Wu, Xiaoyang
Guo, Yuan-Chen
Zhang, Song-Hai
Zhao, Hengshuang
He, Tong
Liu, Xihui
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Utilizing pre-trained 2D large-scale generative models, recent works are capable of generating high-quality novel views from a single in-the-wild image. However, due to the lack of information from multiple views, these works encounter difficulties in generating controllable novel views. In this paper, we present DreamComposer, a flexible and scalable framework that can enhance existing view-aware diffusion models by injecting multi-view conditions. Specifically, DreamComposer first uses a view-aware 3D lifting module to obtain 3D representations of an object from multiple views. Then, it renders the latent features of the target view from 3D representations with the multi-view feature fusion module. Finally the target view features extracted from multi-view inputs are injected into a pre-trained diffusion model. Experiments show that DreamComposer is compatible with state-of-the-art diffusion models for zero-shot novel view synthesis, further enhancing them to generate high-fidelity novel view images with multi-view conditions, ready for controllable 3D object reconstruction and various other applications.
title DreamComposer: Controllable 3D Object Generation via Multi-View Conditions
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2312.03611