Skip to content

Dataset · 1aurent

ADE20K image segmentation dataset

1aurent/ADE20K

A set of over 27,000 annotated images for object segmentation tasks.

What's inside

Contains over 27K images from SUN and Places databases with complete object annotations covering over 3K categories. Includes original polygons, object instances for amodal segmentation, and object parts. Faces and license plates in the images are blurred for anonymization.

What you can do with it

  • Training semantic and instance image segmentation models
  • Developing computer vision systems for recognizing objects and their parts
Tasks
Image segmentation
Languages
English
Size
10K–100K rows
Downloads this month
25.5K
Likes
19

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("1aurent/ADE20K")