---
language: en
license: cc-by-4.0
task_categories:
- image-classification
tags:
- tinyML
- open-images-v6
- object-classification
dataset_info:
  features:
  - name: image
    dtype: image
  - name: label
    dtype:
      class_label:
        names:
          '0': Bicycle&Motorcycle
          '1': Person
          '2': Chair
          '3': Cat
          '4': Houseplant
          '5': Table
          '6': Bottle
          '7': Dog
          '8': Glasses
          '9': Car
  splits:
  - name: calib_images/*
    num_examples: 640
  - name: test_images/*
    num_examples: 50
pretty_name: OpenImagesV6_small
---
# Dataset Card for [OpenImagesV6_small]
<!-- Provide a quick 1-2 sentence summary of what your version of the dataset does -->
This dataset is a custom downstream variant derived from the official [Open Images V6 Dataset](https://storage.googleapis.com/openimages/web/factsfigures.html). 
It has been extracted for object classification applications in TinyML applications. 
## Dataset Description
- **Curated by:** [Omnivision Inc]
- **Source Dataset:** [Open Images Dataset V6](https://storage.googleapis.com/openimages/web/factsfigures.html)
- **Language:** English
- **License:** CC BY 4.0 (for annotations) / Flickr Terms of Use (for source images)
## Licensing and Attribution Information ⚠️
This dataset features a split-licensing structure inheriting from its upstream source, Open Images V6:
1. **Annotations:** Any new or modified annotations, labels, and bounding boxes created for this project are released under the [Creative Commons Attribution 4.0 International (CC BY 4.0)](https://creativecommons.org) license.
2. **Source Images:** The underlying photographs are originating from Flickr. The creators of this dataset do not hold the copyright to these images. Use of the images is bound by the [Flickr Terms of Use](https://flickr.com). Users are individually responsible for ensuring their compliance with the original photographers' terms, particularly regarding commercial downstream use.
### Upstream Attribution
If you use this dataset, you must attribute both this derivative version and the original foundation dataset:
* **Original Dataset Creators:** Open Images ([OpenImages: A public dataset for large-scale multi-label and multi-class image classification. (A. Kuzenetsova et al., 2017)](https://arxiv.org/abs/1912.03098)).
## Citation
If you use this dataset, please cite the following original paper alongside this repository:
```bibtex
@article{OpenImagesV4
  title={The Open Images Dataset V4: Unified image classification, object detection, and visual relationship detection at scale},
  author={A. Kuznetsova, H. Rom, N. Alldrin, J. Uijlings, I. Krasin, J. Pont-Tuset, S. Kamali, S. Popov, M. Malloci, A. Kolesnikov, T. Duerig, and V. Ferrari},
  journal={International Journal of Computer Vision (IJCV)},
  year={2020}
}

@article{InstanceSegment
  title={Large-scale interactive object segmentation with human annotators},
  author={R. Benenson, S. Popov, and V. Ferrari},
  journal={Computer Vision and Pattern Recognition (CVPR)},
  year={2019}
}

@article{LocalizedNarratives
  title={Connecting Vision and Language with Localized Narratives},
  author={J. Pont-Tuset, J. Uijlings, S. Changpinyo, R. Soricut, and V. Ferrari},
  journal={European Conference on Computer Vision (ECCV)},
  year={2020}
}

@article{OpenImages
  title={OpenImages: A public dataset for large-scale multi-label and multi-class image classification},
  author={Krasin I., Duerig T., Alldrin N., Ferrari V., Abu-El-Haija S., Kuznetsova A., Rom H., Uijlings J., Popov S., Kamali S., Malloci M., Pont-Tuset J., Veit A., Belongie S., Gomes V., Gupta A., Sun C., Chechik G., Cai D., Feng Z., Narayanan D., Murphy K.},
  year={2017}
}
```