This project lists representative papers/codes/datasets about Text-Driven 360-Degree Panorama Generation, which aims to comprehensively and systematically summarize the recent advances from 2022 to 2025 to the best of our knowledge.
We aim to constantly update the latest relevant papers and help the community track this topic. If you find any missed resources or errors, please feel free to open an issue or make a pull request here.
Text-only generation focuses on synthesizing 360-degree panoramas from textual descriptions only.
DiffPano: Scalable and Consistent Text to Panorama
Generation with Spherical Epipolar-Aware Diffusion.
Weicai Ye, Chenhao Ji, Zheng Chen, Junyao Gao, Xiaoshui Huang,
Song-Hai Zhang, Wanli Ouyang, Tong He, Cairong Zhao, Guofeng
Zhang.
NeurIPS 2024. [PDF] [Project] [Code]
PanoFree: Tuning-Free Holistic Multi-view Image
Generation with Cross-view Self-Guidance.
Aoming Liu,
Zhong Li, Zhang Chen, Nannan Li, Yi Xu, Bryan A. Plummer.
ECCV
2024. [PDF] [Project] [Code]
Taming Stable Diffusion for Text to 360° Panorama Image
Generation.
Cheng Zhang, Qianyi Wu, Camilo Cruz
Gambardella, Xiaoshui Huang, Dinh Phung, Wanli Ouyang, Jianfei
Cai.
CVPR 2024. [PDF] [Project]
[Code]
Customizing 360-degree panoramas through text-to-image
diffusion models.
Hai Wang, Xiaoyu Xiang, Yuchen Fan,
Jing-Hao Xue.
WACV 2024. [PDF] [Project]
[Code]
Text2Light: Zero-Shot Text-Driven HDR Panorama
Generation.
Zhaoxi Chen, Guangcong Wang, Ziwei
Liu.
TOG 2022 (SIGGRAPH Asia). [PDF] [Project]
[Code]
Diffusion360: Seamless 360 Degree Panoramic Image
Generation based on Diffusion Models.
Mengyang Feng,
Jinlin Liu, Miaomiao Cui, Xuansong Xie.
arxiv 2023. [PDF] [Code]
Text-driven narrow field-of-view (NFoV) outpainting enhances user control by conditioning the generation process on both textual prompts and initial narrow NFoV images.
CubeDiff: Repurposing Diffusion-Based Image Models for
Panorama Generation.
Nikolai Kalischek, Michael
Oechsle, Fabian Manhardt, Philipp Henzler, Konrad Schindler, Federico
Tombari.
ICLR 2025. [PDF] [Project]
Autoregressive Omni-Aware Outpainting for Open-Vocabulary
360-Degree Image Generation.
Zhuqiang Lu, Kun Hu,
Chaoyue Wang, Lei Bai, Zhiyong Wang.
AAAI 2024. [PDF] [Code]
360-Degree Panorama Generation from Few Unregistered NFoV
Images.
Jionghao Wang, Ziyu Chen, Jun Ling, Rong Xie,
Li Song.
ACM MM 2023. [PDF] [Code]
Guided Co-Modulated GAN for 360° Field of View
Extrapolation.
Mohammad Reza Karimi Dastjerdi, Yannick
Hold-Geoffroy, Jonathan Eisenmann, Siavash Khodadadeh, Jean-François
Lalonde.
3DV 2022. [PDF] [Project]
OPa-Ma: Text Guided Mamba for 360-degree Image
Out-painting.
Penglei Gao, Kai Yao, Tiandi Ye, Steven
Wang, Yuan Yao, Xiaofeng Wang.
arxiv 2024. [PDF] [Code]
Diffusion360: Seamless 360 Degree Panoramic Image
Generation based on Diffusion Models.
Mengyang Feng,
Jinlin Liu, Miaomiao Cui, Xuansong Xie.
arxiv 2024. [PDF] [Code]
Recent text-driven 360-degree 3D scene generation methods utilize 360-degree panorama generation to bridge the gap between text prompts and 360-degree 3D scene reconstruction.
DreamScene360: Unconstrained Text-to-3D Scene Generation
with Panoramic Gaussian Splatting.
Shijie Zhou, Zhiwen
Fan, Dejia Xu, Haoran Chang, Pradyumna Chari, Tejas Bharadwaj, Suya You,
Zhangyang Wang, Achuta Kadambi.
ECCV 2024. [PDF] [Project] [Code]
FastScene: Text-Driven Fast 3D Indoor Scene Generation
via Panoramic Gaussian Splatting.
Yikun Ma, Dandan
Zhan, Zhi Jin.
IJCAI 2024. [PDF] [Code]
SceneDreamer360: Text-Driven 3D-Consistent Scene
Generation with Panoramic Gaussian Splatting.
Wenrui
Li, Yapeng Mi, Fucheng Cai, Zhe Yang, Wangmeng Zuo, Xingtao Wang,
Xiaopeng Fan.
arxiv 2024. [PDF] [Project] [Code]
LayerPano3D: Layered 3D Panorama for Hyper-Immersive
Scene Generation.
Shuai Yang, Jing Tan, Mengchen Zhang,
Tong Wu, Yixuan Li, Gordon Wetzstein, Ziwei Liu, Dahua Lin.
arxiv 2024. [PDF] [Project] [Code]
HoloDreamer: Holistic 3D Panoramic World Generation from
Text Descriptions.
Haiyang Zhou, Xinhua Cheng, Wangbo
Yu, Yonghong Tian, Li Yuan.
arxiv 2024. [PDF] [Project] [Code]
Quantitative Comparison of Representative Text-Driven 360-Degree Panorama Generation. We employ an out-of-domain dataset, ODI-SR, on which none of the models have been explicitly trained. Metrics are based on evaluation criteria. Inference time is for generating a 1024×512 panorama. The best and second-best results are highlighted.
| Method | FID ↓ | KID (×10⁻²) ↓ | IS ↑ | CS ↑ | FAED ↓ | OmniFID ↓ | DS ↓ | Inference (s) |
|---|---|---|---|---|---|---|---|---|
| Text-Only Generation | ||||||||
| Text2Light | 72.63 | 1.54 | 5.35 | 19.20 | 18.10 | 99.81 | 5.38 | 33 |
| Diffusion360 | 70.32 | 2.00 | 5.29 | 18.74 | 12.43 | 92.23 | 0.94 | 3 |
| StitchDiffusion | 76.69 | 2.04 | 7.36 | 19.20 | 15.58 | 108.63 | 1.07 | 28 |
| PanFusion | 61.23 | 1.07 | 6.16 | 18.96 | 13.16 | 92.22 | 0.85 | 30 |
| Text-Driven NFoV Outpainting | ||||||||
| PanoDiff | 65.94 | 2.44 | 4.72 | 19.02 | 10.24 | 122.30 | 1.10 | 48 |
| Diffusion360 | 64.19 | 2.05 | 4.53 | 17.92 | 5.50 | 101.39 | 0.72 | 4 |
Summary of popular datasets used in text-driven 360-degree panorama generation. Categories are indoor (I), outdoor (O), or hybrid (I, O).
| Dataset | Year | Category | # Samples | Resolution | License |
|---|---|---|---|---|---|
| SUN360 | 2012 | I & O | 67,583 | 9104 × 4552 | Custom |
| Matterport3D | 2017 | I | 10,800 | 2048 x 1024 | Custom |
| Laval Indoor | 2017 | I | 2,233 | 7668 × 3884 | Custom |
| Laval Outdoor | 2019 | O | 205 | 7668 × 3884 | Custom |
| Structured3D | 2020 | I | 196,515 | 1024 × 512 | Custom |
| Pano360 | 2021 | I & O | 35,000 | 8192 × 4096 | Custom |
| Polyhaven | 2025 | I & O | 786 | 8192 × 4096 | CC0 |
| Humus | 2025 | I & O | 139 | 8192 × 4096 | CC BY 3.0 |
Fréchet Inception Distance (FID) code
Kernel Inception Distance (KID) code
Inception Score (IS) code
CLIP Score (CS) code