Humbi Behavioral Dataset

Opening A New Opportunity For Human Behavioral Imaging

About Us

HUMBI is a large multiview image dataset of human body expressions (gaze, face, hand, body, and garment) with natural clothing. 107 synchronized HD cameras are used to capture more than 700 subjects across gender, ethnicity, age, and style. With the multiview image streams, it provides 3D mesh models. HUMBI is highly effective in learning and reconstructing a complete human model and is complementary to the existing datasets of human body expressions with limited views and subjects.

Software: Access Here

Recent News

August 2021
Raw point clouds with semantic labels are available in Humbi Body.

August 2021
Our new benchmark challenge for pose-guided human rendering is opened.

August 2020
We provide the script to download the dataset per subject for stable access.

June 2020
Our paper is presented in CVPR 2020.

June 2020
The datasets for all body elements and the utility codes are available. For missing subjects, we will keep fixing and updating.

March 2020
HUMBI dataset webpage is open.


[2] HUMBI: A Large Multiview Dataset of Human Body Expressions [paper]

J. S. Yoon,  Z. Yu, J. Park,  and H. S. Park, TPAMI (under review) 

[2] HUMBI: A Large Multiview Dataset of Human Body Expressions [paper] [video]

*Z. Yu, *J. S. Yoon, I. K. Lee, P. Venkatesh, J. Park, J. Yu, and H. S. Park CVPR 2020