Anny-One Dataset
The Anny-One dataset consists of 780K+ multi-person and multi-view synthetic images.
Extreme camera positioning, different cameras, positions and fields of view (FOV)
Download the Anny-One dataset
Anny-One consists of 780,964 RGB images depicting 22,951 synthetic scenes. For each scene, we provide multiple views annotated with camera parameters, human 2D segmentation and 3D pose and shape parameters.
We also provide a Python script demonstrating how to parse the annotations:
An example of Anny-One annotations
News
– 18 Feb 2026: Added segmentation masks and filtered out scenes with annotation issues https://download.europe.naverlabs.com/humans/AnnyOne/scenes_with_issues.txt
– 10 Feb 2026: Initial release.
| AnnyOne_LICENSE.txt | 3.7K |
| AnnyOne_part_1.tar | 17G |
| AnnyOne_part_2.tar | 17G |
| AnnyOne_part_3.tar | 16G |
| AnnyOne_part_4.tar | 16G |
| AnnyOne_part_5.tar | 16G |
| AnnyOne_part_6.tar | 15G |
| AnnyOne_part_7.tar | 12G |
| AnnyOne_part_8.tar | 12G |
| annyone_dataset_visualization.py | 5.0K |
| fast_0xa9327f01_Dini..> | 97K |
| fast_0xa9327f01_Dini..> | 32K |
| fast_0xa9327f01_Dini..> | 8.3K |
| md5sum.txt | 371 |
Reference
Brégier, Romain, et al. “Human Mesh Modeling for Anny Body.” arXiv preprint arXiv:2511.03589 (2025).
Citation
@misc{brégier2025humanmeshmodelinganny,
title={Human Mesh Modeling for Anny Body},
author={Romain Brégier and Guénolé Fiche and Laura Bravo-Sánchez and Thomas Lucas and Matthieu Armando and Philippe Weinzaepfel and Grégory Rogez and Fabien Baradel},
year={2025},
eprint={2511.03589},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2511.03589},
}