| CARVIEW |
UnrealZoo: Enriching Photo-realistic Virtual Worlds for Embodied AI
ICCV 2025 (Highlight)
†BIGAI, ≬City University of Macau, ◊National University of Singapore
* Core Contributors
Abstract
We introduce UnrealZoo, a rich collection of photo-realistic 3D virtual worlds built on Unreal Engine, designed to reflect the complexity and variability of the open worlds. Additionally, we offer a variety of playable entities for embodied AI.
- Photo-realistic Rendering: Empowered by Unreal Engine(4.27/5.4).
- Large-scale Maps: 100+ Maps, the largest one covers 16km².
- Diverse Scenes: Landscape🏜️🏞️🏝️, historic architecture⛩️🛕🏰, settlement🏘️🏙️, industrial zone🏭🏗️, facilities🤽🏻♂️🚉🏪...
- Diverse Bodies: Human🚶🏻♂️️🚶🏻, Robot🤖, Vehicles🚗🏍️🛩️, Animals🐘🐕🐎🐧🐢🐖🐏...
- Diverse Actions: Running🏃🏻♂️, climbing🧗🏻, sitting🧘🏻, jumping, squatting, picking up...
- Easy-to-use: Pre-build binaries with UnrealCV are to be used without any prior knowledge of UE.
- Flexible Python APIs: Provide Python APIs and Gym Interfaces for various potential applications.
- Multi-agent Interactions: Real-time interaction of 10+ vision-based agents👩👩👧👧 in one scene.
Video
Enriching Scenes
The scenes are sourced from the Unreal Engine Marketplace Please refer to the Scene Gallery for more details.
Enriching Interactions
BibTeX
@article{zhong2024unrealzoo,
author = {Zhong, Fangwei and Wu, Kui and Wang, Churan and Chen, Hao and Ci, Hai and Li, Zhoujun and Wang, Yizhou},
title = {UnrealZoo: Enriching Photo-realistic Virtual Worlds for Embodied AI},
journal = {arXiv preprint arXiv:2412.20977},
year = {2024}
}Contact
If you have any suggestion or interested in using or contributing to UnrealZoo, please contact fangweizhong@bnu.edu.cn
Acknowledgments
We would like to thank the Unreal Engine community for providing the engine and the Unreal Engine Marketplace and its contributors for the sharing assets. We also thank Weichao Qiu and other active developers for providing valuable supports in the improvements of UnrealCV plugin.