DreamCube: 3D Panorama Generation via Multi-plane Synchronization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Yukun, Zhou, Yanning, Wang, Jianan, Huang, Kaiyi, Liu, Xihui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912443032141824
author Huang, Yukun
Zhou, Yanning
Wang, Jianan
Huang, Kaiyi
Liu, Xihui
author_facet Huang, Yukun
Zhou, Yanning
Wang, Jianan
Huang, Kaiyi
Liu, Xihui
contents 3D panorama synthesis is a promising yet challenging task that demands high-quality and diverse visual appearance and geometry of the generated omnidirectional content. Existing methods leverage rich image priors from pre-trained 2D foundation models to circumvent the scarcity of 3D panoramic data, but the incompatibility between 3D panoramas and 2D single views limits their effectiveness. In this work, we demonstrate that by applying multi-plane synchronization to the operators from 2D foundation models, their capabilities can be seamlessly extended to the omnidirectional domain. Based on this design, we further introduce DreamCube, a multi-plane RGB-D diffusion model for 3D panorama generation, which maximizes the reuse of 2D foundation model priors to achieve diverse appearances and accurate geometry while maintaining multi-view consistency. Extensive experiments demonstrate the effectiveness of our approach in panoramic image generation, panoramic depth estimation, and 3D scene generation.
format Preprint
id arxiv_https___arxiv_org_abs_2506_17206
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DreamCube: 3D Panorama Generation via Multi-plane Synchronization
Huang, Yukun
Zhou, Yanning
Wang, Jianan
Huang, Kaiyi
Liu, Xihui
Graphics
Computer Vision and Pattern Recognition
Machine Learning
3D panorama synthesis is a promising yet challenging task that demands high-quality and diverse visual appearance and geometry of the generated omnidirectional content. Existing methods leverage rich image priors from pre-trained 2D foundation models to circumvent the scarcity of 3D panoramic data, but the incompatibility between 3D panoramas and 2D single views limits their effectiveness. In this work, we demonstrate that by applying multi-plane synchronization to the operators from 2D foundation models, their capabilities can be seamlessly extended to the omnidirectional domain. Based on this design, we further introduce DreamCube, a multi-plane RGB-D diffusion model for 3D panorama generation, which maximizes the reuse of 2D foundation model priors to achieve diverse appearances and accurate geometry while maintaining multi-view consistency. Extensive experiments demonstrate the effectiveness of our approach in panoramic image generation, panoramic depth estimation, and 3D scene generation.
title DreamCube: 3D Panorama Generation via Multi-plane Synchronization
topic Graphics
Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2506.17206