# cleanlab/cleanvision

Automatically find issues in image datasets and practice data-centric computer vision.

Repository: https://github.com/cleanlab/cleanvision
Canonical: https://ross.abutalabs.com/products/cleanvision
Homepage: https://cleanvision.readthedocs.io/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: computer-vision, data-centric-ai, data-exploration, data-quality, data-validation, deep-learning, exploratory-data-analysis, image-analysis, image-classification, image-generation, image-quality, image-segmentation, data-profiling, data-science
Last push: 2026-01-08T07:19:58+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 61, release rhythm 32, longevity 100
- inputs: {"age_days": 1560, "days_push": 237, "days_rel": 240, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1199, forks 83 (observed 2026-08-28T04:03:57.834183+00:00)

## What it is
CleanVision is a Python library that automatically detects issues in image datasets, such as blurry, dark, over-exposed, or near-duplicate images. It provides a simple API to audit any image collection and generate a report before training machine learning models.

## Use cases
- find blurry or low-quality images in my dataset
- detect near-duplicate images before training a model
- audit an image dataset for data quality issues
- check for dark or over-exposed images in a computer vision dataset
- clean up image data before training a classifier
- profile an image dataset for exploratory data analysis

## When to choose
- you want a quick, automated first-pass quality check of an image dataset before ML training
- you need to find duplicates or low-quality images across a large folder of images with minimal code
- you follow a data-centric AI workflow and want dataset issue reports

## When to avoid
- you need image annotation or labeling tools
- you need general-purpose image editing or manipulation
- your data is not images (e.g., text or tabular data)

## Facets
- artifact type: library
- maturity: active
- function: image-processing, computer-vision, data-science, machine-learning
- domain: computer-vision, image-processing, data-science, machine-learning
- platform: python, cross-platform
- tags: data-quality, data-centric-ai, image-deduplication, dataset-audit, exploratory-data-analysis

## Member repositories
- cleanlab/cleanvision (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:57.834183+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 2026-08-30T06:20:52.642628+00:00, confidence not recorded.
  - readme: https://github.com/cleanlab/cleanvision (fetched 2026-08-28T04:03:57.834183+00:00, sha 87df2512d5d7)
  - registry_pypi: https://pypi.org/pypi/cleanvision/json (fetched 2026-08-29T12:28:57.456836+00:00, sha 22dc7448ee46)
- Data as of 2026-08-30T08:39:29.467469+00:00.