mingyang-wu/Objectron-Videos

# Objectron Videos Mirror

This repository is a videos-only mirror of the official Objectron dataset, prepared for hosting on Hugging Face.

Official source repository:
- https://github.com/google-research-datasets/Objectron

## Purpose

- Provide a clean and upload-friendly copy of Objectron video files.
- Keep directory layout aligned with official dataset conventions.
- Simplify distribution for downstream training and research workflows.

## What Is Included

This mirror currently contains only video files.

Included:
- `videos/<class>/batch-<i>/<j>/<video>.MOV`

Not included in this mirror:
- annotation protobufs (for example `geometry.pbdata`)
- AR metadata protobufs
- tf.records / sequence examples
- index files and train/test split files
- parsing/evaluation scripts

For full dataset assets and tooling, use the official repository and storage paths.

## Directory Layout

The video files follow the official Objectron layout pattern:

- `videos/class/batch-i/j/video.MOV`

Current class folders may include:
- `bike`
- `book`
- `bottle`
- `camera`
- `cereal_box`
- `chair`
- `cup`
- `laptop`
- `shoe`

## License

This repository follows the official Objectron licensing terms.

Objectron is released under:
- Computational Use of Data Agreement 1.0 (C-UDA-1.0)
- https://github.com/microsoft/Computational-Use-of-Data-Agreement

A copy of the license is included in [LICENSE](LICENSE).

## Attribution

If you use Objectron data, please cite the official Objectron paper and follow attribution guidance from the official repository:
- https://github.com/google-research-datasets/Objectron

## Acknowledgment

We thank the Objectron team and the official maintainers for providing this dataset and related resources. These contributions were instrumental in the successful completion of our work: [ConsID-Gen](https://mingyang.me/ConsID-Gen/).

Objectron is a large-scale, object-centric video dataset with pose annotations and has made important contributions to 3D understanding and related vision research.

This repository is only a videos-only mirror for easier access and distribution.

## Disclaimer

- This repository is not an official Google release.
- We cannot guarantee that the number of videos in this mirror exactly matches the counts reported in the original Objectron paper or official storage.
- The contents here only include video files available from our local download process.

## Citation

If you found the original Objectron dataset useful, please cite the official paper.

```bibtex
@article{objectron2021,
title={Objectron: A Large Scale Dataset of Object-Centric Videos in the Wild with Pose Annotations},
author={Adel Ahmadyan, Liangkai Zhang, Artsiom Ablavatski, Jianing Wei, Matthias Grundmann},
journal={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
year={2021}
}
```

This is not an officially supported Google product. If you have any question, you can email us at [e-mail] or join our mailing list at [e-mail].

```bibtex
@misc{wu2026considgenviewconsistentidentitypreservingimagetovideo,
title={ConsID-Gen: View-Consistent and Identity-Preserving Image-to-Video Generation},
author={Mingyang Wu and Ashirbad Mishra and Soumik Dey and Shuo Xing and Naveen Ravipati and Hansi Wu and Binbin Li and Zhengzhong Tu},
year={2026},
eprint={2602.10113},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2602.10113},
}
```