Skip to content

Experience FiftyOne

Sign in with LinkedIn, GitHub, or Google and instantly experience the power of FiftyOne.

Explore more than 50 datasets containing almost 1 million images with detections, embeddings, predictions, keypoints, and segmentations.

Let's go!

Dataset Groups

We’ve tagged the datasets you can try out so it’s easy to find the ones most relevant to your industry or use case. You can also browse by media type like image, video, or point cloud. Here are some popular categories to get you started.

Generative AI
Video
Hugging Face
Automotive
Retail
CVPR + NeurIPS
3D / Point cloud
Keypoints
Segmentation

Available Datasets

array(13) { [0]=> string(9) "thumbnail" [1]=> string(6) "medium" [2]=> string(12) "medium_large" [3]=> string(5) "large" [4]=> string(9) "1536x1536" [5]=> string(9) "2048x2048" [6]=> string(18) "Image Size 500x500" [7]=> string(28) "gb-block-post-grid-landscape" [8]=> string(25) "gb-block-post-grid-square" [9]=> string(23) "card_thumb_center_small" [10]=> string(25) "card_thumb_top_crop_small" [11]=> string(17) "card_thumb_center" [12]=> string(19) "card_thumb_top_crop" }

  • ARKitTrack

    A diverse dataset for tracking
using Mobile RGB-D data

  • ARMBench-ObjectID-100

    A subset of Amazon's object-centric benchmark dataset for robotic manipulation

  • ASL_Dataset

    A real-life large-scale sign language
dataset comprising over 25,000
annotated videos

  • BDD100k

    Large driving video dataset with
100K videos

  • BDD100K-Validation

    Subset of the large driving video dataset with 100K videos

  • Brian-Thesis

    PDF document loaded as images for OCR and document search

  • Building3D

    An urban-scale dataset for 3D building reconstruction

  • CaFFe

    Synthetic Aperture Radar (SAR) images of seven glaciers distributed around the globe

  • Caltech Fish Counting

    Large-scale dataset for detecting, tracking, and counting fish in sonar videos

  • ChestX-ray14

    Medical imaging dataset which comprises 112,120 frontal-view X-ray images

  • Clothing Segmentation

    Clothing Co-Parsing dataset of high-resolution street fashion photos

  • COCO 2017 Validation

    Subset of the large-scale object detection, segmentation & captioning dataset

  • COCO Demo

    Subset of the large-scale object detection, segmentation & captioning dataset

  • dacl10k

    A benchmark for semantic bridge damage segmentation

  • EgoObjects-Val

    Large-scale egocentric dataset for fine-grained object understanding

  • Emojis

    High-res images of ~2,000 emojis. Click the emoji button in the menu to semantically search through emojis!

  • EqBen-Test

    Benchmark VLMs with visual-minimal change samples

  • Example-Groups

    A subset of the KITTI dataset showcasing grouped media (images & LIDAR) for each scene

  • Example-Images

    The best starting point for exploring the FiftyOne App!

  • Example-Videos

    A small video dataset of 10 video segments with dense object detections

  • Fashion Product Images

    44k products with multiple category labels, descriptions & high-res images

  • Football Player Segmentation

    Player detection & segmentation in football matches

  • Ford AV Data

    Challenging multi-agent seasonal dataset collected by Ford autonomous vehicles

  • GeoNet

    Large-scale, curated dataset of images from the USA and Asia

  • ImageNet-E

    A rigorous benchmark for evaluating image classifier robustness

  • ImageRewardDB

    Comparison dataset for evaluating human text-to-image preferences

  • JRDB

    A novel dataset collected from the JackRabbot social mobile manipulator

  • JRDB-Pose

    One of the largest datasets of ground truth human body pose annotations

  • KITTI

    A popular dataset for mobile robotics and autonomous driving

  • KITTI-Multiview-Train

    A subset of the popular dataset for mobile robotics and autonomous driving

  • LLaVA-Instruct

    A set of GPT-generated multimodal instruction-following data

  • LLCM

    A dataset constructed to facilitate the study of low-light cross-modality person Re-ID tasks

  • Magic Brush

    A manually annotated dataset for instruction-guided image editing

  • Mobile-HDR

    Dataset of phone cameras paired to collect LDR-HDR images

  • MOSE

    Dataset for video object segmentation in complex scenes

  • MVImgNet

    A large-scale dataset of multi-view images

  • NFL Impact Detection

    Dataset to detect helmet impacts in videos of NFL plays

  • NIGHTS

    Synthetic dataset of image triplets with human scores of perceptual similarity

  • nuScenes

    A large-scale autonomous driving dataset with 3D object annotations

  • Open Images V7

    A dataset of images with labels spanning thousands of object categories

  • PedX

    Benchmark dataset for metric 3D pose estimation of pedestrians in complex urban intersections

  • Quickstart Groups Demo

    Demo dataset of multiple slices of multimodal samples organized into groups

  • Real3D-AD

    A dataset of point cloud anomaly detection

  • Retail Product Checkout

    Large-scale and fine-grained retail product checkout dataset

  • Satlas Marine Infrastructure

    Geospatial data generated by AI from satellite imagery of offshore wind turbines & platforms

  • SoccerNet-v3

    A large-scale dataset for soccer video understanding

  • SportsMOT

    A large multi-object tracking dataset in multiple sports scenes

  • Spring

    A large, high-resolution, computer-generated benchmark for scene flow, optical flow & stereo

  • SynSL-120K

    Synthetic data with blender for bare supervised learning and downstream domain adaptation

  • Wildlife Watcher

    A dataset of images used in wildlife conservation

  • WildMe

    Animal re-identification dataset spanning 3 species & various viewpoints, curated by the nonprofit Wild Me
Load more

Submit a Dataset

Do you have an interesting dataset you’d like to include in FiftyOne to get increased exposure, plus make it easy to browse?

Include My Dataset

Ready for a FiftyOne Demo?

Request a FiftyOne Teams demo to learn more about how plugins, data versioning, cloud-backed media and enterprise-grade security can turbocharge your team’s AI and computer vision workflows.