Skip to content

Dataset · detection-datasets

Fashionpedia: fashion image annotation

detection-datasets/fashionpedia

Fashion object detection dataset: images with segmentation masks and expert clothing ontology.

What's inside

Fashionpedia consists of an ontology from fashion experts and an image dataset. The ontology has 27 main clothing categories, 19 clothing parts, 294 fine-grained attributes, and relationships between them. Casual and celebrity fashion images are annotated with segmentation masks; size is 10k to 100k examples, English language.

What you can do with it

  • Training object detection and clothing segmentation
  • Testing models on fine-grained fashion object attributes
  • Creating a detector for annotating fashion catalogs
Tasks
Object detection
Languages
English
Size
10K–100K rows
Downloads this month
4.5K
Likes
87

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("detection-datasets/fashionpedia")