IDEA-Research/HumanArt

[CVPR 2023] The official implementation of CVPR 2023 paper "Human-Art: A Versatile Human-Centric Dataset Bridging Natural and Artificial Scenes"

32
/ 100
Emerging

This project provides a large dataset, Human-Art, containing 50,000 images across 20 diverse scenarios, featuring both real and artificial human figures. Each image includes detailed annotations for human bounding boxes, 2D keypoints, self-contact keypoints, and descriptive text. It's designed for researchers and developers working on computer vision tasks like detecting humans, estimating 2D and 3D human poses, and generating realistic or stylized human images.

275 stars. No commits in the last 6 months.

Use this if you need a comprehensive, multi-scenario dataset to train and evaluate models for human detection, pose estimation, or controllable human image generation across various contexts, including out-of-domain scenarios.

Not ideal if your project focuses solely on highly specific niche applications that do not involve general human detection or pose understanding.

computer-vision human-detection pose-estimation image-generation dataset-development
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 6 / 25

How are scores calculated?

Stars

275

Forks

4

Language

License

Apache-2.0

Last pushed

Oct 03, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/IDEA-Research/HumanArt"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.