# openimages/dataset

The Open Images dataset

Repository: https://github.com/openimages/dataset
Canonical: https://ross.abutalabs.com/products/openimages-dataset
Homepage: https://storage.googleapis.com/openimages/web/index.html
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2021-07-01T18:02:46+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3625, "days_push": 1889, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4376, forks 606 (observed 2026-08-28T04:08:46.907615+00:00)

## What it is
Open Images is a large-scale annotated image dataset for computer vision, containing millions of images with bounding boxes, instance segmentations, visual relationships, localized narratives, and image-level labels across thousands of classes. This repository serves as the dataset's home, with the dataset itself hosted on a dedicated site for browsing and downloading.

## Use cases
- train an object detection model on a large labeled image dataset
- find images with instance segmentation masks for semantic segmentation research
- get image-level labels for weakly supervised learning
- download annotated images for visual relationship detection
- benchmark computer vision models on a standard dataset
- find images with point-level annotations for zero-shot classification

## When to choose
- you need a large, diverse, freely licensed image dataset with rich annotations
- you are training or benchmarking object detection, segmentation, or classification models
- you need thousands of object classes beyond what smaller datasets offer

## When to avoid
- you need a small, lightweight dataset for quick experiments
- you need video or audio data rather than still images
- you cannot handle multi-gigabyte downloads or need fully curated, clean labels out of the box

## Facets
- artifact type: dataset
- maturity: maintenance
- function: computer-vision, image-processing, machine-learning
- domain: computer-vision, machine-learning, image-processing, data-science
- platform: python, cross-platform
- tags: image-dataset, object-detection, image-segmentation, visual-relationship-detection, annotations, computer-vision-benchmark

## Member repositories
- openimages/dataset (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:46.907615+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:21:20.111236+00:00, confidence not recorded.
  - readme: https://github.com/openimages/dataset (fetched 2026-08-28T04:08:46.907615+00:00, sha 413065e1b44e)
  - homepage: https://storage.googleapis.com/openimages/web/index.html (fetched 2026-08-29T09:09:37.493163+00:00, sha c26c27290afc)
- Data as of 2026-08-30T08:39:29.467469+00:00.