Taming Stable Diffusion for Text to 360° Panorama Image Generation

CVPR 2024 Hightlight

Cheng Zhang 1,3     Qianyi Wu 1     Camilo Cruz Gambardella 1,3     Xiaoshui Huang 2
Dinh Phung 1     Wanli Ouyang 2     Jianfei Cai 1
1 Monash University     2 Shanghai AI Laboratory     3 Building 4.0 CRC, Caulfield East, Victoria, Australia
Cite Code arXiv Paper

Abstract

Generative models, e.g., Stable Diffusion, have enabled the creation of photorealistic images from text prompts. Yet, the generation of 360-degree panorama images from text remains a challenge, particularly due to the dearth of paired text-panorama data and the domain gap between panorama and perspective images. In this paper, we introduce a novel dual-branch diffusion model named PanFusion to generate a 360-degree image from a text prompt. We leverage the stable diffusion model as one branch to provide prior knowledge in natural image generation and register it to another panorama branch for holistic image generation. We propose a unique cross-attention mechanism with projection awareness to minimize distortion during the collaborative denoising process. Our experiments validate that PanFusion surpasses existing methods and, thanks to its dual-branch structure, can integrate additional constraints like room layout for customized panorama outputs. Code is available at https://chengzhag.github.io/publication/panfusion.

Pipeline

Our proposed dual-branch PanFusion pipeline. The panorama branch (upper) provides global layout guidance and registers the perspective information to get seamless panorama output. The perspective branch (lower) harnesses the rich prior knowledge of Stable Diffusion (SD) and provides guidance to alleviate distortion under perspective projection. Both branches employ the same UNet backbone with shared weights, while finetuned with separate LoRA layers. Equirectangular-Perspective Projection Attention (EPPA) modules are plugged into different layers of the UNet to pass information between the two branches.
Our proposed dual-branch PanFusion pipeline. The panorama branch (upper) provides global layout guidance and registers the perspective information to get seamless panorama output. The perspective branch (lower) harnesses the rich prior knowledge of Stable Diffusion (SD) and provides guidance to alleviate distortion under perspective projection. Both branches employ the same UNet backbone with shared weights, while finetuned with separate LoRA layers. Equirectangular-Perspective Projection Attention (EPPA) modules are plugged into different layers of the UNet to pass information between the two branches.

Comparisons

629_supp_Page_05.jpg 629_supp_Page_06.jpg 629_supp_Page_07.jpg 629_supp_Page_08.jpg 629_supp_Page_09.jpg 629_supp_Page_10.jpg 629_supp_Page_11.jpg 629_supp_Page_12.jpg 629_supp_Page_13.jpg 629_supp_Page_14.jpg

Generalization

629_supp_Page_17.jpg 629_supp_Page_18.jpg 629_supp_Page_19.jpg 629_supp_Page_20.jpg
Deep Learning Image Generation Panorama
Cheng Zhang
Cheng Zhang
Ph.D Student

My research interests include 3D generation, novel view synthesis, image generation etc.