---
title: "Best Free Computer Vision Datasets: The Ultimate Guide for AI Developers in 2026"
description: "Free computer vision datasets differ less in size than in license: MS COCO and Open Images annotations are CC BY 4.0, while ImageNet, Cityscapes, CelebA,…"
url: "https://unidata.pro/blog/best-free-computer-vision-datasets/"
date_modified: "2026-10-27T08:53:26+03:00"
language: "en-US"
---
![Best Free Computer Vision Datasets: The Ultimate Guide for AI Developers in 2026](https://unidata.pro/wp-content/uploads/2026/10/computer-vision-datasets.webp)Free computer vision datasets differ less in size than in license: MS COCO and Open Images annotations are CC BY 4.0, while ImageNet, Cityscapes, CelebA, DOTA and Virtual KITTI 2 are free only for non-commercial research. Check that line before anything else if the model may ship.

Below are 13 datasets with volumes and access terms checked against official pages and papers in September 2026, grouped by task.

## Foundation datasets for classification and detection

### 1. [ImageNet](https://www.image-net.org/)

![Imagenet screenshot dataset](https://unidata.pro/wp-content/uploads/2026/10/1.-imagenet.webp)**Volume:** 14.2 million images across 21,841 WordNet synsets; the ILSVRC-2012 subset has 1,281,167 training images in 1,000 classes[1](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-1)
**Access:** Free with login and agreement to terms; non-commercial research and education only
**Format:** JPEG images organized by WordNet hierarchy
**Task fit:** Image classification, localization, transfer learning, backbone pretraining

The benchmark every new backbone is compared against, and the source of most pretrained weights in TorchVision and Hugging Face. Almost all practical work uses the 1,000-class ILSVRC subset rather than the full set.

### 2. [Open Images Dataset V7](https://storage.googleapis.com/openimages/web/index.html)

![Open Images Dataset V7 screenshot](https://unidata.pro/wp-content/uploads/2026/10/2.-open-images-dataset-v7.webp)**Volume:** 9.18 million images; 15,851,536 boxes across 600 classes; 2,785,498 instance masks across 350 classes; 66.4 million point labels[2](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-2)
**Access:** Free, no signup; annotations CC BY 4.0, images listed as CC BY 2.0 (check each image's license before commercial use)
**Format:** Images + CSV annotations (boxes, masks, points, relationships)
**Task fit:** Object detection, instance segmentation, visual relationship detection

Flickr-sourced images give a real-world long tail, which makes it a good generalization check once a model scores well on COCO. Use FiftyOne or the official downloader to pull a class subset instead of the full set.

### 3. [MS COCO](https://cocodataset.org/)

![MS COCO screenshot](https://unidata.pro/wp-content/uploads/2026/10/3.-ms-coco.webp)**Volume:** 328,000 images with 2.5 million labeled instances, 80 object categories, keypoints and five captions per image[3](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-3); the 91-class "stuff" layer comes from COCO-Stuff[4](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-4)
**Access:** Free, open download; annotations CC BY 4.0, images under Flickr terms
**Format:** Images + JSON annotations (COCO format)
**Task fit:** Object detection, instance and panoptic segmentation, keypoints, captioning

The default comparison point for objects in context: papers report its mAP, and nearly every toolkit reads its [annotation format](https://cocodataset.org/#format-data). Fewer classes than Open Images, but more consistent labels.

### 4. [PASCAL VOC 2012](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/)

![PASCAL VOC 2012 screenshot](https://unidata.pro/wp-content/uploads/2026/10/4.-pascal-voc-2012.webp)**Volume:** 11,530 train/val images, 20 classes, 27,450 annotated objects, 6,929 segmentations; test set evaluated on a submission server[5](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-5)
**Access:** Free download of train/val; no explicit dataset license, images under Flickr terms
**Format:** Images + XML annotations
**Task fit:** Object detection, semantic segmentation, action classification

Small enough for a same-day training run and still in the baseline tables of current papers. Treat it as a development set, not a primary training source.

## Segmentation and scene understanding

### 5. [Cityscapes](https://www.cityscapes-dataset.com/)

![Cityscapes screenshot](https://unidata.pro/wp-content/uploads/2026/10/5.-cityscapes.webp)**Volume:** 5,000 finely annotated and 20,000 coarsely annotated images, 30 classes, street scenes from 50 cities[6](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-6)
**Access:** Free registration for people affiliated with a research body; [non-commercial license](https://www.cityscapes-dataset.com/license/), no redistribution
**Format:** Images + pixel-level label PNGs
**Task fit:** Semantic, instance and panoptic segmentation

The standard for urban semantic segmentation. The coarse subset trades precision for volume, so pick the split your loss actually needs.

### 6. [ADE20K](https://ade20k.csail.mit.edu/)

![ADE20K](https://unidata.pro/wp-content/uploads/2026/10/6.-ade20k.webp)**Volume:** Full set of 25,574 training and 2,000 validation images; the common "20K images, 150 classes" figure is the SceneParse150 benchmark subset (20,210 training images)[7](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-7)
**Access:** Full set after registration, images for non-commercial research and education only; SceneParse150 is a direct download; annotations BSD-3
**Format:** Images + pixel labels
**Task fit:** Semantic segmentation, scene parsing, open-vocabulary segmentation

Covers indoor and outdoor scenes, which Cityscapes does not. Check whether a paper reports on the full set or SceneParse150 before comparing numbers.

### 7. [SA-1B (Segment Anything 1 Billion)](https://ai.meta.com/datasets/segment-anything/)

![SA-1B (Segment Anything 1 Billion) screenshot](https://unidata.pro/wp-content/uploads/2026/10/7.-sa-1b-segment-anything.webp)**Volume:** 11 million images, 1.1 billion masks, about 100 masks per image[8](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-8)
**Access:** Free for research use under the SA-1B Dataset Research License
**Format:** Images + RLE-encoded masks (COCO-style)
**Task fit:** Promptable and zero-shot segmentation, foundation-model training

The data behind the [Segment Anything Model](https://github.com/facebookresearch/segment-anything). The full release runs to several terabytes, so most teams use a subset or pretrained SAM weights.

## Video and face datasets

### 8. [Kinetics-700](https://github.com/cvdfoundation/kinetics-dataset)

![Kinetics-700 dataset](https://unidata.pro/wp-content/uploads/2026/10/8.-kinetics-700.webp)**Volume:** 700 action classes with at least 600 clips each in the 2019 release, about 650,000 clips[10](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-10); Kinetics-700-2020 raises the floor to 700 clips per class
**Access:** Free, no signup, via the CVDF mirror; no license statement found, clips come from YouTube
**Format:** Short video clips, one action label each
**Task fit:** Action recognition, video classification, temporal modeling

The default action-recognition benchmark[9](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-9). Clips disappear from YouTube over time, so the usable set shrinks.

### 9. [CelebA](https://mmlab.ie.cuhk.edu.hk/projects/CelebA.html)

![CelebA](https://unidata.pro/wp-content/uploads/2026/10/9.-celeba.webp)**Volume:** 202,599 face images of 10,177 identities, 40 binary attributes and 5 landmarks per image[11](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-11)
**Access:** Free for non-commercial research only; no redistribution
**Format:** Images + attribute and landmark annotations
**Task fit:** Face attribute recognition, landmark detection, generative modeling

Still common in GAN and attribute-learning papers. Celebrity photography skews demographics and pose, so audit for [bias across groups](https://unidata.pro/blog/understanding-bias-in-ml/) before building on it.

### 10. [LFW (Labeled Faces in the Wild)](https://vis-www.cs.umass.edu/lfw/)

![LFW (Labeled Faces in the Wild) screeshot](https://unidata.pro/wp-content/uploads/2026/10/10.-lfw.webp)**Volume:** 13,233 images of 5,749 people, 1,680 of them with two or more images[12](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-12)
**Access:** Free download; the authors caution that LFW scores do not show an algorithm is suitable for commercial use
**Format:** Images + match/mismatch pair labels
**Task fit:** Face verification in unconstrained conditions

A quick sanity check for a verification pipeline. Current models saturate it, so a strong score is a floor, not a result.

## Domain-specific and synthetic datasets

### 11. [DOTA-v2.0](https://captain-whu.github.io/DOTA/)

![DOTA-v2.0 screenshot](https://unidata.pro/wp-content/uploads/2026/10/11.-dota-v2.0.webp)**Volume:** 11,268 aerial images, 1,793,658 instances, 18 categories, oriented bounding boxes[13](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-13)
**Access:** Free for academic use; commercial use prohibited
**Format:** Images + oriented bounding-box annotations
**Task fit:** Oriented object detection, aerial and remote-sensing imagery

The reference set for drone and satellite detection, where objects appear at any rotation. Geography and sensors are fixed, so test on your own region before trusting results.

### 12. [Mapillary Vistas](https://www.mapillary.com/dataset/vistas)

![Mapillary Vistas](https://unidata.pro/wp-content/uploads/2026/10/12.-mapillary-vistas.webp)**Volume:** 25,000 images (18,000 train, 2,000 validation, 5,000 test), 66 categories, 37 with instance labels[14](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-14)
**Access:** Free with signup for the research edition; a separate commercial edition exists
**Format:** Images + segmentation labels
**Task fit:** Street-scene segmentation, cross-dataset generalization

Crowdsourced images from many cities and cameras. Pair it with Cityscapes: Cityscapes for label density, Mapillary for how a model holds up elsewhere.

### 13. [Virtual KITTI 2](https://europe.naverlabs.com/proxy-virtual-worlds-vkitti-2/)

**Volume:** Five KITTI tracking sequences re-rendered as clone, fog, morning, overcast, rain and sunset, plus 15° and 30° left and right camera rotations[15](https://unidata.pro/blog/best-free-computer-vision-datasets/#ref-15)
**Access:** Free, non-commercial only (CC BY-NC-SA 3.0)
**Format:** RGB + depth + segmentation + optical flow + camera pose
**Task fit:** Domain adaptation, synthetic-to-real transfer, multi-task learning

Isolates domain shift on fixed scene geometry rather than generating new content. See our [synthetic data guide](https://unidata.pro/blog/synthetic-data/) for when that is enough.

## Quick comparison table

| **Dataset** | **Category** | **Access** | **Format** | **Task fit** |
|---|---|---|---|---|
| ImageNet | Classification | Non-commercial, login | JPEG + WordNet labels | Classification, pretraining |
| Open Images V7 | Detection / segmentation | CC BY 4.0 annotations | CSV boxes + masks | Detection, relationships |
| MS COCO | Detection / segmentation | CC BY 4.0 annotations | COCO JSON | Detection, captioning, keypoints |
| PASCAL VOC 2012 | Detection / segmentation | Free, no explicit license | XML | Baselines |
| Cityscapes | Segmentation | Non-commercial, registration | Pixel-level labels | Urban segmentation |
| ADE20K | Segmentation | Non-commercial images | Pixel labels | Scene parsing |
| SA-1B | Segmentation | Research, Meta license | RLE masks | Promptable segmentation |
| Kinetics-700 | Video / action | Free, no license stated | Video clips + labels | Action recognition |
| CelebA | Face | Non-commercial | Images + attributes | Attributes, GANs |
| LFW | Face | Free | Images + pair labels | Face verification |
| DOTA-v2.0 | Aerial | Academic only | Oriented boxes | Aerial detection |
| Mapillary Vistas | Street scenes | Research edition, signup | Segmentation labels | Generalization |
| Virtual KITTI 2 | Synthetic | Non-commercial | RGB + depth + flow | Domain adaptation |

## How to choose a computer vision dataset

Check three things, in this order:

1. **Task:** classification, detection, segmentation, video or faces. This picks the section above.
2. **Format:** COCO JSON is read by Detectron2, MMDetection, Ultralytics and FiftyOne; VOC XML and [YOLO text labels](https://docs.ultralytics.com/datasets/detect/) are also common. Decide your target format before downloading.
3. **License:** only Open Images and COCO annotations are CC BY 4.0; most others above are non-commercial. Resolve this before a model reaches a product decision.

## When the free datasets run out

Public datasets are for prototyping and benchmarking. Production models usually need data from the client's own objects, environments and edge cases. Dense aerial scenes show the gap first: on an [aerial imagery annotation project](https://unidata.pro/cases/aerial-image-annotation-for-urban-planning/), frames with thousands of objects slowed the tool, and assigning annotators by object class kept the work moving. On a [license-plate annotation project](https://unidata.pro/cases/license-plate-annotation-for-vehicle-recognition-system/), the annotated real data outperformed datasets built with synthetic plates. For the decision itself, see [build or buy an ML dataset](https://unidata.pro/blog/build-or-buy-ml-dataset/).

If your project has outgrown the free datasets and needs annotation built around your own images — [start here](https://unidata.pro/data-annotation/image/).

## Additional Modules

### FAQ Section

**Questions and Answers:**

- **Question:** What are the best free computer vision datasets for object detection? — **Answer:** MS COCO and Open Images V7 cover the most ground, with large scale, rich annotations and near-universal tooling. PASCAL VOC is fastest for baselines, and DOTA-v2.0 is the choice for oriented or aerial boxes.
- **Question:** Which free computer vision datasets can be used commercially? — **Answer:** Open Images and MS COCO annotations are released under CC BY 4.0, but their images carry Flickr licenses that must be checked individually. ImageNet, Cityscapes, CelebA, DOTA-v2.0, Virtual KITTI 2 and ADE20K images are non-commercial, and SA-1B is research-only. For a commercial product, plan on licensed or custom-collected data.
- **Question:** Which datasets are best for semantic segmentation? — **Answer:** Cityscapes for urban scenes, ADE20K or its SceneParse150 subset for general scene parsing, SA-1B for promptable segmentation at scale, and Mapillary Vistas for geographic diversity. Combining two is standard for generalization testing.
- **Question:** Where can I download free computer vision datasets? — **Answer:** Most are hosted on their official project pages, linked in each heading above. Hugging Face Datasets and FiftyOne load many of them in one call. Always confirm the license before commercial use.
- **Question:** Are synthetic datasets like Virtual KITTI 2 still useful? — **Answer:** Yes, for testing domain adaptation and generating conditions such as fog or unusual camera angles that are expensive to capture. They do not replace real data once a task depends on real-world noise, so pair them with a real validation set.

### references

**List of Links:**

- **Link:** [[1] Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., Fei-Fei, L. "ImageNet: A Large-Scale Hierarchical Image Database" — CVPR — 2009](https://www.image-net.org/static_files/papers/imagenet_cvpr09.pdf) — **Active link:** Active link
- **Link:** [[2] Kuznetsova, A., Rom, H., Alldrin, N., et al. "The Open Images Dataset V4: Unified Image Classification, Object Detection, and Visual Relationship Detection at Scale" — IJCV — 2020](https://arxiv.org/abs/1811.00982) — **Active link:** Active link
- **Link:** [[3] Lin, T.-Y., Maire, M., Belongie, S., et al. "Microsoft COCO: Common Objects in Context" — ECCV — 2014](https://arxiv.org/abs/1405.0312) — **Active link:** Active link
- **Link:** [[4] Caesar, H., Uijlings, J., Ferrari, V. "COCO-Stuff: Thing and Stuff Classes in Context" — CVPR — 2018](https://arxiv.org/abs/1612.03716) — **Active link:** Active link
- **Link:** [[5] Everingham, M., Van Gool, L., Williams, C.K.I., Winn, J., Zisserman, A. "The PASCAL Visual Object Classes Challenge: A Retrospective" — IJCV — 2015](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/) — **Active link:** Active link
- **Link:** [[6] Cordts, M., Omran, M., Ramos, S., et al. "The Cityscapes Dataset for Semantic Urban Scene Understanding" — CVPR — 2016](https://arxiv.org/abs/1604.01685) — **Active link:** Active link
- **Link:** [[7] Zhou, B., Zhao, H., Puig, X., Fidler, S., Barriuso, A., Torralba, A. "Scene Parsing through ADE20K Dataset" — CVPR — 2017 ](https://ade20k.csail.mit.edu/) — **Active link:** Active link
- **Link:** [[8] Kirillov, A., Mintun, E., Ravi, N., et al. "Segment Anything" — ICCV — 2023](https://arxiv.org/abs/2304.02643) — **Active link:** Active link
- **Link:** [[9] Kay, W., Carreira, J., Simonyan, K., et al. "The Kinetics Human Action Video Dataset" — arXiv preprint — 2017 ](https://arxiv.org/abs/1705.06950) — **Active link:** Active link
- **Link:** [[10] Carreira, J., Noland, E., Hillier, C., Zisserman, A. "A Short Note on the Kinetics-700 Human Action Dataset" — arXiv preprint — 2019](https://arxiv.org/abs/1907.06987) — **Active link:** Active link
- **Link:** [[11] Liu, Z., Luo, P., Wang, X., Tang, X. "Deep Learning Face Attributes in the Wild" — ICCV — 2015 ](https://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) — **Active link:** Active link
- **Link:** [[12] Huang, G.B., Ramesh, M., Berg, T., Learned-Miller, E. "Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments" — University of Massachusetts, Amherst, Technical Report 07-49 — 2007](https://vis-www.cs.umass.edu/lfw/) — **Active link:** Active link
- **Link:** [[13] Ding, J., Xue, N., Xia, G.-S., et al. "Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges" — IEEE TPAMI — 2021](https://arxiv.org/abs/2102.12219) — **Active link:** Active link
- **Link:** [[14] Neuhold, G., Ollmann, T., Rota Bulò, S., Kontschieder, P. "The Mapillary Vistas Dataset for Semantic Understanding of Street Scenes" — ICCV — 2017](https://openaccess.thecvf.com/content_ICCV_2017/papers/Neuhold_The_Mapillary_Vistas_ICCV_2017_paper.pdf) — **Active link:** Active link
- **Link:** [[15] Cabon, Y., Murray, N., Humenberger, M. "Virtual KITTI 2" — arXiv preprint — 2020](https://arxiv.org/abs/2001.10773) — **Active link:** Active link

[Full list of this site's AI-readable pages](https://unidata.pro/llms.txt)